> For the complete documentation index, see [llms.txt](https://docs-conversational-ai.syntphony.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs-conversational-ai.syntphony.com/api-docs/appendices/environment-data-structure.md).

# Environment Data Structure

Understand how Syntphony Conversational AI stores the data, configuration, and relationships available inside an environment.

Each environment uses an isolated database structure. The schema contains the resources created and used by Conversational Projects, including channels, conversations, NLP resources, Knowledge AI, Agentic capabilities, integrations, voice configuration, and system data.

Use this page when you need to identify:

* Where a resource is stored
* Which fields are available
* Whether a field is required
* How records connect to one another
* Which data can be queried from the replica database.

## Understand the environment schema

The schema is built around a few central relationships:

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th></tr></thead><tbody><tr><td>Area</td><td>Main tables</td><td>Purpose</td></tr><tr><td>Conversational Project</td><td><code>bot</code>, <code>configuration</code>, <code>tags</code></td><td>Stores project-level identity, configuration, and reusable metadata.</td></tr><tr><td>Agentic capabilities</td><td><code>agent</code>, <code>persona</code>, <code>agent_function</code>, <code>agent_rule</code>, <code>agent_knowledge</code>, <code>mcp_server</code></td><td>Defines agent behavior, identity, functions, knowledge access, and external tools.</td></tr><tr><td>Knowledge and generative AI</td><td><code>collection</code>, <code>document</code>, <code>question</code>, <code>generative_service</code></td><td>Stores Knowledge AI resources, Prompt Cells, and training data.</td></tr><tr><td>Channels and voice</td><td><code>channel</code>, <code>channel_type</code>, <code>channel_ivr_configuration</code>, <code>channel_dnis_ivr_configuration</code></td><td>Defines channel configuration, voice behavior, and DNIS associations.</td></tr><tr><td>NLP</td><td><code>nlp_engine</code>, <code>intents</code>, <code>utterance</code>, <code>entity</code></td><td>Stores cognitive engines, intents, utterances, and entities.</td></tr><tr><td>Conversations</td><td><code>session</code>, <code>user_interaction</code>, <code>satisfaction</code></td><td>Stores runtime conversation and user interaction data.</td></tr><tr><td>Validation</td><td><code>automated_tests</code>, <code>training</code>, <code>training_al</code></td><td>Stores testing and training information.</td></tr></tbody></table>

A `bot` record represents the internal database object for a Conversational Project. Many tables use `bot_uuid` to associate data with that project.

## Environment schema

### Agent

The `agent` table stores agents configured in a Conversational Project.

An agent can coordinate the next step as a supervisor or interact directly with the user. Its record connects behavior, identity, model settings, execution limits, and the persona used during the conversation.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the agent.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this agent belongs to.</td></tr><tr><td><code>persona_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the persona that defines the agent's identity.</td></tr><tr><td><code>type</code></td><td>varchar(36)</td><td>Yes</td><td>Agent type, such as supervisor or agent.</td></tr><tr><td><code>role</code></td><td>varchar(50)</td><td>No</td><td>Role assigned to the agent, for example <code>support</code>.</td></tr><tr><td><code>goal</code></td><td>text</td><td>No</td><td>Outcome this agent is expected to achieve.</td></tr><tr><td><code>instructions</code></td><td>text</td><td>No</td><td>Instructions that guide how the agent behaves and responds.</td></tr><tr><td><code>guardrails</code></td><td>text</td><td>No</td><td>Boundaries that limit what the agent is allowed to do.</td></tr><tr><td><code>fallback</code></td><td>varchar(2000)</td><td>No</td><td>Behavior followed by the supervisor when it cannot handle the user's request.</td></tr><tr><td><code>maximum_tokens</code></td><td>int(11)</td><td>No</td><td>Maximum number of tokens allowed in one response.</td></tr><tr><td><code>request_timeout</code></td><td>bigint</td><td>No</td><td>Time, in seconds, the agent waits for the model before the request fails.</td></tr><tr><td><code>temperature</code></td><td>float(7,4)</td><td>No</td><td>Controls response variability for GPT-4 models.</td></tr><tr><td><code>topp</code></td><td>float(7,4)</td><td>No</td><td>Controls the range of likely tokens available to GPT-4 models.</td></tr><tr><td><code>presence_penalty</code></td><td>float(7,4)</td><td>No</td><td>Discourages GPT-4 models from introducing new topics.</td></tr><tr><td><code>frequency_penalty</code></td><td>float(7,4)</td><td>No</td><td>Discourages GPT-4 models from repeating the same words.</td></tr><tr><td><code>verbosity</code></td><td>varchar(20)</td><td>No</td><td>Response length for GPT-5 models. Values can include <code>LOW</code>, <code>MEDIUM</code>, or <code>HIGH</code>.</td></tr><tr><td><code>reasoning_effort</code></td><td>varchar(20)</td><td>No</td><td>Reasoning level for GPT-5 models. Values can include <code>MINIMAL</code>, <code>LOW</code>, <code>MEDIUM</code>, or <code>HIGH</code>.</td></tr><tr><td><code>inherit_from_supervisor</code></td><td>tinyint(1)</td><td>No</td><td>True when this agent uses settings defined by its supervisor.</td></tr><tr><td><code>enabled</code></td><td>tinyint(1)</td><td>No</td><td>True when the agent is enabled.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>No</td><td>True when the agent was deleted. This may be reverted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Agent Function

The `agent_function` table stores functions that agents can call.

Functions extend agent behavior beyond conversation generation. They can request data from another system, perform an operation, or return control to another part of the Agentic flow.

The same function **can be assigned to more than one agent.**

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the function.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this function belongs to.</td></tr><tr><td><code>name</code></td><td>varchar(50)</td><td>Yes</td><td>Function name.</td></tr><tr><td><code>description</code></td><td>text</td><td>Yes</td><td>Explains what the function does so an agent can determine when to call it.</td></tr><tr><td><code>parameter_json</code></td><td>text</td><td>No</td><td>JSON describing the parameters expected by the function.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>No</td><td>True when the function was deleted. This may be reverted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Agent Function Parameter

The `agent_function_parameter` table stores individual inputs accepted by an agent function.

Each parameter defines the value expected by the function and can optionally connect that value to a variable used during execution.

The same parameter name **cannot be repeated within one function**.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the parameter.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this parameter belongs to.</td></tr><tr><td><code>agent_function_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the function this parameter belongs to.</td></tr><tr><td><code>name</code></td><td>varchar(50)</td><td>Yes</td><td>Parameter name.</td></tr><tr><td><code>type</code></td><td>varchar(30)</td><td>Yes</td><td>Expected value type, such as string, integer, or boolean.</td></tr><tr><td><code>description</code></td><td>text</td><td>Yes</td><td>Explains what this parameter represents.</td></tr><tr><td><code>variable</code></td><td>varchar(50)</td><td>No</td><td>Variable populated with this parameter during execution.</td></tr><tr><td><code>optional</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the function can be called without this parameter.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Agent Function Parameter Rule

The `agent_function_parameter_rule` table stores allowed or expected values for a function parameter.

Rules make parameter behavior more explicit by defining a value or constraint accepted by the function.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the rule.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this rule belongs to.</td></tr><tr><td><code>agent_function_parameter_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the parameter this rule applies to.</td></tr><tr><td><code>value</code></td><td>varchar(150)</td><td>Yes</td><td>Value or constraint applied to the parameter.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Agent Knowledge

The `agent_knowledge` table connects an agent to the Knowledge AI collections it can consult.

This relationship controls which knowledge sources are available to the agent. The same collection cannot be associated twice with the same agent.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing this relationship.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this relationship belongs to.</td></tr><tr><td><code>collection_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the collection available to the agent.</td></tr><tr><td><code>agent_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the agent that uses the collection.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Agent Rule

The `agent_rule` table stores rules associated with an agent.

A rule restricts values or patterns the agent can accept. The same rule value cannot be repeated for the same agent within the same Conversational Project.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the rule.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this rule belongs to.</td></tr><tr><td><code>agent_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the agent this rule applies to.</td></tr><tr><td><code>value</code></td><td>varchar(150)</td><td>Yes</td><td>Rule value, such as a keyword or pattern.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Answer

The `answer` table stores response configurations created in a Conversational Project.

An answer can include its presentation data, webhook behavior, authentication, masking, transaction settings, and model configuration for Rephrase Answer.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the answer.</td></tr><tr><td><code>name</code></td><td>varchar(255)</td><td>Yes</td><td>Answer name. Visible only in the Cockpit.</td></tr><tr><td><code>description</code></td><td>text</td><td>No</td><td>Answer description.</td></tr><tr><td><code>auth_type</code></td><td>enum</td><td>Yes</td><td>Authentication used when the answer calls a webhook. Values can include <code>AUTH_NONE</code>, <code>AUTH_BASIC</code>, <code>AUTH_BEARER</code>, or <code>AUTH_OAUTH</code>.</td></tr><tr><td><code>basic_token</code></td><td>varchar(2600)</td><td>No</td><td>Token used with <code>AUTH_BASIC</code>.</td></tr><tr><td><code>bearer_token</code></td><td>varchar(2600)</td><td>No</td><td>Token used with <code>AUTH_BEARER</code>.</td></tr><tr><td><code>o_auth_url</code></td><td>varchar(500)</td><td>No</td><td>URL used to obtain an OAuth token.</td></tr><tr><td><code>grant_type</code></td><td>enum</td><td>No</td><td>OAuth grant type, such as <code>CLIENT_CREDENTIALS</code>, <code>PASSWORD</code>, or <code>REFRESH_TOKEN</code>.</td></tr><tr><td><code>client_id</code></td><td>varchar(150)</td><td>No</td><td>Client ID used with <code>CLIENT_CREDENTIALS</code>.</td></tr><tr><td><code>client_secret</code></td><td>varchar(150)</td><td>No</td><td>Client secret used with <code>CLIENT_CREDENTIALS</code>.</td></tr><tr><td><code>auth_username</code></td><td>varchar(150)</td><td>No</td><td>Username used with the <code>PASSWORD</code> grant type.</td></tr><tr><td><code>auth_password</code></td><td>varchar(150)</td><td>No</td><td>Password used with the <code>PASSWORD</code> grant type.</td></tr><tr><td><code>evaluable</code></td><td>tinyint(1)</td><td>No</td><td>True when the message can be evaluated by the user.</td></tr><tr><td><code>transactional</code></td><td>tinyint(1)</td><td>No</td><td>True when the response is transactional.</td></tr><tr><td><code>is_masked</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the answer content must be masked.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the answer was deleted. This may be reverted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this answer belongs to.</td></tr><tr><td><code>webhook</code></td><td>varchar(255)</td><td>No</td><td>API URL called when the answer is executed.</td></tr><tr><td><code>type</code></td><td>varchar(20)</td><td>Yes</td><td>Answer type. The usual value is <code>ANSWER</code>.</td></tr><tr><td><code>headers</code></td><td>longtext</td><td>No</td><td>Freeform headers sent to the webhook.</td></tr><tr><td><code>execution_limit</code></td><td>int(3)</td><td>No</td><td>Maximum number of consecutive executions. <code>0</code> means no limit.</td></tr><tr><td><code>rephrase_enabled</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the answer is rephrased before being sent.</td></tr><tr><td><code>rephrasing_temperature</code></td><td>float(7,4)</td><td>Yes</td><td>Controls how much rephrased content can vary from the original.</td></tr><tr><td><code>rephrasing_memory</code></td><td>int(11)</td><td>Yes</td><td>Number of previous messages considered during rephrasing.</td></tr><tr><td><code>rephrasing_blacklist</code></td><td>varchar(1000)</td><td>Yes</td><td>Words that must not appear in the rephrased response.</td></tr><tr><td><code>rephrasing_verbosity</code></td><td>varchar(20)</td><td>No</td><td>Response length used when rephrasing with GPT-5 models.</td></tr><tr><td><code>rephrasing_reasoning_effort</code></td><td>varchar(20)</td><td>No</td><td>Reasoning effort used when rephrasing with GPT-5 models.</td></tr></tbody></table>

### Answer Template

The `answer_template` table stores channel-specific content for an answer or question.

An answer can have a different template for each channel. A record points to either an answer or a question.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the answer template.</td></tr><tr><td><code>content</code></td><td>text</td><td>No</td><td>JSON representing the selected template.</td></tr><tr><td><code>answer_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the answer this template belongs to. Empty when associated with a question.</td></tr><tr><td><code>channel_type_id</code></td><td>bigint(20)</td><td>No</td><td>ID of the associated channel type.</td></tr><tr><td><code>question_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the question this template belongs to. Empty when associated with an answer.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr><tr><td><code>type</code></td><td>varchar(20)</td><td>Yes</td><td>Selected template type.</td></tr><tr><td><code>channel_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the channel this template refers to.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the template was deleted.</td></tr><tr><td><code>technical_text</code></td><td>text</td><td>No</td><td>Technical text stored as JSON.</td></tr><tr><td><code>dynamic_content</code></td><td>tinyint(1)</td><td>No</td><td>True when the template content is built dynamically during the conversation.</td></tr></tbody></table>

### API Key

The `api_key` table stores API keys used to validate Syntphony CAI service calls.

| Column Name    | Data Type   | Not Null | Description               |
| -------------- | ----------- | -------- | ------------------------- |
| `api_key`      | varchar(50) | Yes      | API key value.            |
| `created_date` | datetime    | Yes      | Date the key was created. |

### Automated Test

The `automated_tests` table stores the definition of an automated test.

Test cases are provided through the Cockpit. Execution data is stored separately so repeated runs do not overwrite the test definition.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Automated test UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project being tested.</td></tr><tr><td><code>channel_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the channel used by the test.</td></tr><tr><td><code>user_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created the test.</td></tr><tr><td><code>name</code></td><td>varchar(200)</td><td>No</td><td>Test name.</td></tr><tr><td><code>utterances</code></td><td>int(11)</td><td>No</td><td>Number of examples included in the test.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>No</td><td>True when the test was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr></tbody></table>

### Automated Tests Execution

The `automated_tests_execution` table stores one result for each execution of an automated test.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Execution UUID.</td></tr><tr><td><code>automated_test_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the automated test that produced this result.</td></tr><tr><td><code>avg_score</code></td><td>decimal(12,10)</td><td>Yes</td><td>Average confidence score.</td></tr><tr><td><code>precision_score</code></td><td>decimal(12,10)</td><td>Yes</td><td>Precision score.</td></tr><tr><td><code>status</code></td><td>varchar(10)</td><td>Yes</td><td>Execution status.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this execution was last updated.</td></tr></tbody></table>

### Automated Test Utterance

The `automated_test_utterance` table stores one example included in an automated test.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Example UUID.</td></tr><tr><td><code>automated_test_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the automated test this example belongs to.</td></tr><tr><td><code>description</code></td><td>text</td><td>No</td><td>Description of the example.</td></tr><tr><td><code>utterance</code></td><td>text</td><td>No</td><td>Phrase being tested.</td></tr><tr><td><code>expected_category</code></td><td>varchar(200)</td><td>No</td><td>Expected category.</td></tr><tr><td><code>expected_value</code></td><td>varchar(255)</td><td>No</td><td>Expected value.</td></tr><tr><td><code>expected_answer</code></td><td>text</td><td>No</td><td>Expected textual answer.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>No</td><td>True when the object was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Automated Test Utterance Execution

The `automated_test_utterance_execution` table stores the result produced for each utterance during a test execution.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Execution UUID.</td></tr><tr><td><code>automated_test_utterance_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the utterance being evaluated.</td></tr><tr><td><code>automated_test_execution_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the related automated test execution.</td></tr><tr><td><code>given_category</code></td><td>varchar(200)</td><td>No</td><td>Category returned by the execution.</td></tr><tr><td><code>given_value</code></td><td>text</td><td>No</td><td>Value returned by the execution.</td></tr><tr><td><code>given_answer</code></td><td>text</td><td>No</td><td>Textual answer returned by the execution.</td></tr><tr><td><code>session_code</code></td><td>varchar(200)</td><td>No</td><td>Session associated with this execution.</td></tr><tr><td><code>score</code></td><td>decimal(12,10)</td><td>Yes</td><td>Execution score.</td></tr><tr><td><code>proccess_time</code></td><td>int(11)</td><td>No</td><td>Processing time in milliseconds.</td></tr><tr><td><code>description</code></td><td>text</td><td>No</td><td>Optional description.</td></tr></tbody></table>

## Bot

The `bot` table stores the core data of a Conversational Project.

`bot` is the internal database name. The product-facing concept is a Conversational Project.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project.</td></tr><tr><td><code>name</code></td><td>varchar(50)</td><td>Yes</td><td>Conversational Project name.</td></tr><tr><td><code>locale</code></td><td>varchar(6)</td><td>Yes</td><td>Primary locale.</td></tr><tr><td><code>description</code></td><td>text</td><td>No</td><td>Freeform description.</td></tr><tr><td><code>company_name</code></td><td>varchar(50)</td><td>No</td><td>Company represented by this Conversational Project.</td></tr><tr><td><code>industry</code></td><td>varchar(256)</td><td>No</td><td>Business segment, such as retail or telecommunications.</td></tr><tr><td><code>image</code></td><td>varchar(255)</td><td>No</td><td>Image path or URL.</td></tr><tr><td><code>image_content_type</code></td><td>varchar(50)</td><td>No</td><td>Image content type.</td></tr><tr><td><code>bot_analytic_api_key</code></td><td>varchar(50)</td><td>No</td><td>API key used for analytics tracking.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>No</td><td>True when the object was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Bulk Training

The `bulk_training` table stores bulk imports of intents or utterances and tracks the status of the import.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Bulk training UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project receiving the file.</td></tr><tr><td><code>type</code></td><td>enum(<code>INTENTS</code>,<code>UTTERANCES</code>)</td><td>Yes</td><td>Type of bulk import.</td></tr><tr><td><code>percentage</code></td><td>int(11)</td><td>Yes</td><td>Current execution progress.</td></tr><tr><td><code>status</code></td><td>enum(<code>PENDING</code>,<code>FINISH</code>,<code>ERROR</code>)</td><td>Yes</td><td>Import status.</td></tr><tr><td><code>result</code></td><td>longtext</td><td>No</td><td>Inserted objects.</td></tr><tr><td><code>file_name</code></td><td>text</td><td>No</td><td>Submitted filename.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr></tbody></table>

### Bulk Training Item

The `bulk_training_item` table stores one row from a bulk training file.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Item UUID.</td></tr><tr><td><code>bulk_training_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the bulk training this item belongs to.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project.</td></tr><tr><td><code>intent</code></td><td>varchar(255)</td><td>Yes</td><td>Intent name.</td></tr><tr><td><code>utterance</code></td><td>varchar(1000)</td><td>Yes</td><td>Example utterance associated with the intent.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr></tbody></table>

## Channel

The `channel` table stores channels configured for a Conversational Project.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Channel UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this channel belongs to.</td></tr><tr><td><code>channel_type_id</code></td><td>bigint(20) unsigned</td><td>Yes</td><td>ID of the channel type.</td></tr><tr><td><code>name</code></td><td>varchar(45)</td><td>Yes</td><td>Channel name.</td></tr><tr><td><code>description</code></td><td>text</td><td>No</td><td>Freeform description.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>No</td><td>True when the channel was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Channel Classification

The `channel_classification` table groups channel types into channel families.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>id</code></td><td>bigint(20) unsigned</td><td>Yes</td><td>Channel classification ID.</td></tr><tr><td><code>name</code></td><td>varchar(50)</td><td>Yes</td><td>Classification name.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Channel DNIS IVR Configuration

The `channel_dnis_ivr_configuration` table associates a DNIS with a channel and with a shared IVR configuration.

This structure allows one channel to contain multiple DNIS values while applying one IVR configuration to them. One association can be identified as the default DNIS.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the DNIS association.</td></tr><tr><td><code>channel_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the channel this DNIS belongs to.</td></tr><tr><td><code>channel_ivr_configuration_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the IVR configuration used by this DNIS.</td></tr><tr><td><code>dnis</code></td><td>varchar(13)</td><td>Yes</td><td>Dialed number associated with the channel.</td></tr><tr><td><code>is_default</code></td><td>tinyint(1)</td><td>Yes</td><td>True when this is the default DNIS of the channel.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>Yes</td><td>True when this association was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>Yes</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>Yes</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Channel IVR Configuration

The `channel_ivr_configuration` table stores IVR behavior for a voice channel.

Its primary key is `uuid`, allowing the configuration to be referenced independently from a single DNIS.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the IVR configuration. Primary key.</td></tr><tr><td><code>dnis</code></td><td>varchar(13)</td><td>Yes</td><td>Dialed number associated with the IVR configuration.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project handling the call.</td></tr><tr><td><code>channel_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the channel handling the call.</td></tr><tr><td><code>ivr_vendor</code></td><td>varchar(14)</td><td>Yes</td><td>IVR vendor, such as <code>EVG</code> or <code>CISCO</code>. Default: <code>EVG</code>.</td></tr><tr><td><code>fetch_timeout</code></td><td>bigint</td><td>Yes</td><td>Maximum fetch wait in milliseconds. Default: <code>45000</code>.</td></tr><tr><td><code>fetch_audio</code></td><td>text</td><td>Yes</td><td>Audio played during fetch events.</td></tr><tr><td><code>fetch_audio_delay</code></td><td>bigint</td><td>Yes</td><td>Delay before fetch audio begins.</td></tr><tr><td><code>fetch_audio_minimum</code></td><td>bigint</td><td>Yes</td><td>Minimum playback duration after fetch audio starts.</td></tr><tr><td><code>fetch_audio_interval</code></td><td>bigint</td><td>Yes</td><td>Interval between fetch audio loops. <code>-1</code> prevents looping.</td></tr><tr><td><code>inter_digit_timeout</code></td><td>bigint</td><td>Yes</td><td>Maximum interval between DTMF digits.</td></tr><tr><td><code>num_of_digits</code></td><td>bigint</td><td>Yes</td><td>Number of DTMF digits to capture.</td></tr><tr><td><code>term_timeout</code></td><td>bigint</td><td>Yes</td><td>Time after the last input before input terminates.</td></tr><tr><td><code>voice_timeout</code></td><td>bigint</td><td>Yes</td><td>Timeout for voice input.</td></tr><tr><td><code>term_char</code></td><td>varchar(1)</td><td>Yes</td><td>Character that ends DTMF input.</td></tr><tr><td><code>dtmf_timeout</code></td><td>bigint</td><td>Yes</td><td>Wait for DTMF input.</td></tr><tr><td><code>barge_in</code></td><td>tinyint(1)</td><td>Yes</td><td>True when DTMF can interrupt audio playback.</td></tr><tr><td><code>voice_barge_in</code></td><td>tinyint(1)</td><td>Yes</td><td>True when voice barge-in is enabled.</td></tr><tr><td><code>flush</code></td><td>tinyint(1)</td><td>No</td><td>True when audio is flushed rather than queued.</td></tr><tr><td><code>barge_in_offset</code></td><td>bigint</td><td>No</td><td>Point from which interruption is accepted.</td></tr><tr><td><code>time</code></td><td>bigint</td><td>Yes</td><td>Silence duration in milliseconds.</td></tr><tr><td><code>asr_voice_provider</code></td><td>varchar(25)</td><td>No</td><td>ASR provider.</td></tr><tr><td><code>sensitivity</code></td><td>float(7,4)</td><td>No</td><td>Noise reduction sensitivity.</td></tr><tr><td><code>confidence</code></td><td>float(4,3)</td><td>Yes</td><td>Speech recognition confidence threshold.</td></tr><tr><td><code>max_speech_timeout</code></td><td>bigint</td><td>No</td><td>Maximum duration of user speech.</td></tr><tr><td><code>incomplete_timeout</code></td><td>bigint</td><td>No</td><td>Wait for incomplete speech recognition.</td></tr><tr><td><code>asr_voice_provider_config</code></td><td>text</td><td>No</td><td>Credentials used to access the ASR provider.</td></tr><tr><td><code>tts_voice_provider</code></td><td>varchar(25)</td><td>No</td><td>TTS provider.</td></tr><tr><td><code>mask</code></td><td>text</td><td>No</td><td>SSML mask applied to spoken text.</td></tr><tr><td><code>tts_voice_provider_config</code></td><td>text</td><td>No</td><td>Credentials used to access the TTS provider.</td></tr><tr><td><code>default_error_transfer_on_error</code></td><td>tinyint(1)</td><td>Yes</td><td>True when a default error triggers a transfer.</td></tr><tr><td><code>default_error_audio</code></td><td>text</td><td>No</td><td>Audio URL played for a default error.</td></tr><tr><td><code>default_error_tts</code></td><td>text</td><td>No</td><td>Text or SSML synthesized for a default error.</td></tr><tr><td><code>default_error_uui</code></td><td>text</td><td>No</td><td>UUI sent with a default error transfer.</td></tr><tr><td><code>default_error_dest</code></td><td>text</td><td>No</td><td>Destination used for a default error transfer.</td></tr><tr><td><code>call_establishment_error_audio</code></td><td>text</td><td>No</td><td>Audio used when the call cannot be established.</td></tr><tr><td><code>call_establishment_error_tts</code></td><td>text</td><td>No</td><td>Text or SSML used when the call cannot be established.</td></tr><tr><td><code>call_establishment_error_uui</code></td><td>text</td><td>No</td><td>UUI sent when the call cannot be established.</td></tr><tr><td><code>call_establishment_error_dest</code></td><td>text</td><td>No</td><td>Transfer destination used when the call cannot be established.</td></tr><tr><td><code>call_establishment_error_transfer_on_error</code></td><td>tinyint(1)</td><td>Yes</td><td>True when failed call establishment triggers a transfer.</td></tr><tr><td><code>dest</code></td><td>varchar(256)</td><td>No</td><td>Call transfer destination, such as SIP or telephone.</td></tr><tr><td><code>uui</code></td><td>varchar(256)</td><td>No</td><td>Custom value sent in the SIP user-to-user header.</td></tr><tr><td><code>first_conversation_request</code></td><td>text</td><td>No</td><td>Request used to start the conversation on another flow.</td></tr><tr><td><code>send_post_call_events</code></td><td>tinyint(1)</td><td>Yes</td><td>True when post call events are sent.</td></tr><tr><td><code>regional_expressions_file_url</code></td><td>text</td><td>No</td><td>URL of the JSON file containing regional expressions.</td></tr><tr><td><code>welcome_timeout</code></td><td>bigint</td><td>Yes</td><td>Maximum wait for the welcome flow.</td></tr><tr><td><code>conversation_timeout</code></td><td>bigint</td><td>Yes</td><td>Maximum wait for a conversation response.</td></tr><tr><td><code>max_call_duration</code></td><td>bigint</td><td>Yes</td><td>Maximum duration of an active call.</td></tr><tr><td><code>secret</code></td><td>varchar(16)</td><td>Yes</td><td>Key used to authenticate the initial IVR call.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Channel Type

The `channel_type` table stores channel types and connects each type to its classification.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>id</code></td><td>bigint(20) unsigned</td><td>Yes</td><td>Channel type ID.</td></tr><tr><td><code>channel_classification_id</code></td><td>bigint(20) unsigned</td><td>Yes</td><td>ID of the related channel classification.</td></tr><tr><td><code>name</code></td><td>varchar(50)</td><td>Yes</td><td>Channel type name.</td></tr><tr><td><code>image</code></td><td>varchar(30)</td><td>No</td><td>Image used to represent the channel type.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

## Collection

The `collection` table stores Knowledge AI collections.

Collections group documents by subject and define how those documents participate in retrieval. A Conversational Project can have one default collection.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the collection.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this collection belongs to.</td></tr><tr><td><code>name</code></td><td>varchar(100)</td><td>Yes</td><td>Collection name shown in the Cockpit.</td></tr><tr><td><code>instructions</code></td><td>varchar(5000)</td><td>No</td><td>Instructions describing how this collection should be used.</td></tr><tr><td><code>description</code></td><td>varchar(250)</td><td>No</td><td>Freeform description.</td></tr><tr><td><code>is_default</code></td><td>tinyint(1)</td><td>Yes</td><td>True when this is the default collection.</td></tr><tr><td><code>semantic_weight</code></td><td>tinyint(1)</td><td>Yes</td><td>Weight used when selecting results across multiple collections.</td></tr><tr><td><code>previous_user_input</code></td><td>tinyint(1)</td><td>Yes</td><td>Number of previous user messages considered during retrieval.</td></tr><tr><td><code>threshold</code></td><td>tinyint(1)</td><td>Yes</td><td>Minimum similarity required for a document excerpt to be returned.</td></tr><tr><td><code>top_k</code></td><td>tinyint(1)</td><td>Yes</td><td>Maximum number of excerpts returned.</td></tr><tr><td><code>training_status</code></td><td>varchar(30)</td><td>Yes</td><td>Collection training status.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the collection was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>Yes</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>Yes</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

## Configuration

The `configuration` table stores environment configuration keys.

A record can represent either a parameter or a variable. When `bot_uuid` contains a value, the configuration belongs to a specific Conversational Project. Otherwise it applies to the environment.

Variable records also identify the context and value type used at runtime.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the configuration.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the Conversational Project when the configuration is project-specific.</td></tr><tr><td><code>configuration_type_id</code></td><td>tinyint unsigned</td><td>Yes</td><td>Configuration type ID. <code>1</code> = parameter, <code>2</code> = variable.</td></tr><tr><td><code>key</code></td><td>varchar(50)</td><td>Yes</td><td>Configuration key.</td></tr><tr><td><code>context_type</code></td><td>enum</td><td>No</td><td>Variable context. Values can include <code>hiddenContext</code>, <code>visibleContext</code>, or <code>openContext</code>.</td></tr><tr><td><code>context_key</code></td><td>varchar(50)</td><td>No</td><td>Key used to identify or group the variable inside its context.</td></tr><tr><td><code>value_type</code></td><td>enum</td><td>No</td><td>Stored value type: <code>null</code>, <code>string</code>, <code>number</code>, <code>boolean</code>, or <code>json</code>.</td></tr><tr><td><code>value</code></td><td>varchar(2500)</td><td>No</td><td>Configuration value.</td></tr><tr><td><code>description</code></td><td>varchar(300)</td><td>No</td><td>Configuration description.</td></tr><tr><td><code>enabled</code></td><td>tinyint(1)</td><td>No</td><td>True when the configuration is enabled.</td></tr><tr><td><code>erasable</code></td><td>tinyint(1)</td><td>No</td><td>True when the configuration can be fully erased.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the configuration was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Configuration Type

The `configuration_type` table identifies the kind of record stored in `configuration`.

Initial values are:

* `1` — parameter
* `2` — variable

| Column Name | Data Type        | Not Null | Description                               |
| ----------- | ---------------- | -------- | ----------------------------------------- |
| `id`        | tinyint unsigned | Yes      | Configuration type ID.                    |
| `name`      | varchar(10)      | Yes      | Type name, such as parameter or variable. |

## Document

The `document` table stores documents supplied to Knowledge AI.

Each record identifies the document, its collection, file format, storage location, training state, and the text-window settings used during processing.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the document.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this document belongs to.</td></tr><tr><td><code>collection_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the collection containing this document.</td></tr><tr><td><code>format_id</code></td><td>tinyint unsigned</td><td>Yes</td><td>File-format ID related to <code>document_format</code>.</td></tr><tr><td><code>storage_filename</code></td><td>varchar(42)</td><td>Yes</td><td>Path to the stored document file.</td></tr><tr><td><code>training_status</code></td><td>varchar(30)</td><td>No</td><td>Document training status.</td></tr><tr><td><code>filename</code></td><td>varchar(255)</td><td>Yes</td><td>Document filename.</td></tr><tr><td><code>windows_length</code></td><td>smallint</td><td>No</td><td>Size of each text window generated during training.</td></tr><tr><td><code>overlap</code></td><td>smallint</td><td>No</td><td>Amount of text shared between consecutive windows.</td></tr><tr><td><code>enabled</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the document is enabled.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the document was deleted.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr></tbody></table>

## Entity

The `entity` table stores entities created in a Conversational Project.

An entity defines a type of information the NLP layer can identify in user input.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the entity.</td></tr><tr><td><code>name</code></td><td>varchar(255)</td><td>Yes</td><td>Entity name.</td></tr><tr><td><code>description</code></td><td>varchar(250)</td><td>No</td><td>Freeform description.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this entity belongs to.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the object was deleted.</td></tr><tr><td><code>enabled</code></td><td>tinyint(1)</td><td>No</td><td>True when the object is enabled.</td></tr><tr><td><code>is_masked</code></td><td>tinyint(1)</td><td>Yes</td><td>True when values identified for this entity must be masked.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr><tr><td><code>metadata</code></td><td>varchar(200)</td><td>No</td><td>NLP metadata used when the project relies on an external NLP engine.</td></tr></tbody></table>

### Entity Sample

The `entity_sample` table stores synonyms or patterns associated with an entity value.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Sample UUID.</td></tr><tr><td><code>entity_value_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the entity value this sample belongs to.</td></tr><tr><td><code>value</code></td><td>varchar(255)</td><td>Yes</td><td>Synonym or pattern.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Entity System

The `entity_system` table maps the system entities available to NLP engines.

| Column Name | Data Type    | Not Null | Description         |
| ----------- | ------------ | -------- | ------------------- |
| `id`        | bigint(20)   | Yes      | System entity ID.   |
| `name`      | varchar(255) | Yes      | System entity name. |

### Entity Value

The `entity_value` table stores values belonging to an entity.

For example, an entity named `"sport"` could contain values such as `"football"`, `"basketball"`, or `"tennis"`.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Entity value UUID.</td></tr><tr><td><code>entity_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the entity this value belongs to.</td></tr><tr><td><code>value</code></td><td>varchar(255)</td><td>Yes</td><td>Entity value.</td></tr><tr><td><code>type</code></td><td>varchar(20)</td><td>Yes</td><td>Value type, such as synonym or pattern.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

## Facebook Configuration

The `facebook_configuration` table stores the channel configuration associated with a Facebook page.

| Column Name         | Data Type    | Not Null | Description                                      |
| ------------------- | ------------ | -------- | ------------------------------------------------ |
| `page_id`           | varchar(45)  | No       | Facebook page ID.                                |
| `page_name`         | varchar(45)  | No       | Facebook page name.                              |
| `hub_token`         | varchar(100) | Yes      | Verification token used by the Facebook webhook. |
| `page_access_token` | varchar(225) | No       | Token used to validate page access.              |
| `channel_uuid`      | varchar(36)  | Yes      | UUID of the associated channel.                  |

## Facebook User

The `facebook_user` table stores information about a Facebook user that interacted with the Conversational Project.

| Column Name        | Data Type   | Not Null | Description                                     |
| ------------------ | ----------- | -------- | ----------------------------------------------- |
| `user_id`          | varchar(20) | Yes      | Facebook user ID.                               |
| `create_date`      | datetime    | No       | Record creation date.                           |
| `user_blocked`     | tinyint(2)  | No       | Indicates whether the user is blocked.          |
| `last_interaction` | datetime    | No       | Date and time of the user's latest interaction. |

## Generative Service

The `generative_service` table stores Prompt Cells created in a Conversational Project.

A Prompt Cell sends a prompt to a language model and stores the generated response in a variable so it can be used later in the flow.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID representing the Prompt Cell.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this Prompt Cell belongs to.</td></tr><tr><td><code>name</code></td><td>varchar(50)</td><td>Yes</td><td>Prompt Cell name shown in the Cockpit.</td></tr><tr><td><code>description</code></td><td>text</td><td>No</td><td>Freeform description.</td></tr><tr><td><code>prompt</code></td><td>longtext</td><td>Yes</td><td>Prompt sent to the language model.</td></tr><tr><td><code>model</code></td><td>varchar(25)</td><td>Yes</td><td>Language model used by this Prompt Cell.</td></tr><tr><td><code>verbosity</code></td><td>varchar(20)</td><td>No</td><td>Response length used by GPT-5 models.</td></tr><tr><td><code>reasoning_effort</code></td><td>varchar(20)</td><td>No</td><td>Reasoning level used by GPT-5 models.</td></tr><tr><td><code>temperature</code></td><td>float(7,4)</td><td>No</td><td>Controls response variability for GPT-4 models.</td></tr><tr><td><code>maximum_tokens</code></td><td>int(11)</td><td>Yes</td><td>Maximum number of tokens the model can return.</td></tr><tr><td><code>topp</code></td><td>float(7,4)</td><td>No</td><td>Controls the range of likely tokens available to GPT-4 models.</td></tr><tr><td><code>presence_penalty</code></td><td>float(7,4)</td><td>No</td><td>Discourages introducing new topics.</td></tr><tr><td><code>frequency_penalty</code></td><td>float(7,4)</td><td>No</td><td>Discourages repeated wording.</td></tr><tr><td><code>response_variable_name</code></td><td>varchar(50)</td><td>Yes</td><td>Variable where the generated response is stored.</td></tr><tr><td><code>uses_blacklist</code></td><td>tinyint(1)</td><td>No</td><td>True when blacklist words must be avoided.</td></tr><tr><td><code>blacklist_words</code></td><td>text</td><td>No</td><td>Words that must not appear in the generated response.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>No</td><td>True when the Prompt Cell was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

## Input

The `input` table stores Wait input cells configured in a dialog flow.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Input UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project.</td></tr><tr><td><code>type_id</code></td><td>bigint(20)</td><td>Yes</td><td>ID of the input type.</td></tr><tr><td><code>description</code></td><td>mediumtext</td><td>No</td><td>Freeform description.</td></tr><tr><td><code>pattern</code></td><td>mediumtext</td><td>No</td><td>Input pattern.</td></tr><tr><td><code>call_to_action</code></td><td>mediumtext</td><td>No</td><td>Call-to-action configuration.</td></tr><tr><td><code>remember_input</code></td><td>tinyint(4)</td><td>No</td><td>Indicates whether the input should be stored.</td></tr><tr><td><code>stored</code></td><td>varchar(100)</td><td>No</td><td>Stored value.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Input Type

The `input_type` table lists the available input types.

| Column Name | Data Type    | Not Null | Description      |
| ----------- | ------------ | -------- | ---------------- |
| `id`        | bigint(20)   | Yes      | Input type ID.   |
| `type`      | varchar(100) | Yes      | Input type name. |

## Intents

The `intents` table stores intent configurations created in a Conversational Project.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Intent UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this intent belongs to.</td></tr><tr><td><code>name</code></td><td>varchar(255)</td><td>Yes</td><td>Intent name.</td></tr><tr><td><code>description</code></td><td>text</td><td>No</td><td>Intent description.</td></tr><tr><td><code>metadata</code></td><td>varchar(200)</td><td>No</td><td>Metadata used by supported external NLP engines.</td></tr><tr><td><code>enabled</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the intent is enabled.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the intent was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

## MCP Server

The `mcp_server` table stores an MCP server configured for a Conversational Project.

The server instruction tells the system when that MCP server should be called. Authentication and transport information are stored with the server configuration.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>MCP server UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this server belongs to.</td></tr><tr><td><code>name</code></td><td>varchar(50)</td><td>Yes</td><td>MCP server name shown in the Cockpit.</td></tr><tr><td><code>description</code></td><td>varchar(150)</td><td>No</td><td>Freeform description.</td></tr><tr><td><code>instruction</code></td><td>text</td><td>Yes</td><td>Instruction used to determine when the MCP server should be called.</td></tr><tr><td><code>version</code></td><td>varchar(20)</td><td>No</td><td>MCP server version.</td></tr><tr><td><code>transport_type</code></td><td>varchar(30)</td><td>Yes</td><td>Transport used to communicate with the MCP server.</td></tr><tr><td><code>url</code></td><td>varchar(500)</td><td>Yes</td><td>MCP server endpoint.</td></tr><tr><td><code>headers</code></td><td>longtext</td><td>No</td><td>HTTP headers sent with requests.</td></tr><tr><td><code>auth_type</code></td><td>enum</td><td>Yes</td><td>Authentication method used when calling the server.</td></tr><tr><td><code>basic_token</code></td><td>longtext</td><td>No</td><td>Token used with basic authentication.</td></tr><tr><td><code>bearer_token</code></td><td>longtext</td><td>No</td><td>Bearer token.</td></tr><tr><td><code>o_auth_url</code></td><td>varchar(500)</td><td>No</td><td>OAuth token endpoint.</td></tr><tr><td><code>grant_type</code></td><td>enum</td><td>No</td><td>OAuth grant type.</td></tr><tr><td><code>client_id</code></td><td>varchar(150)</td><td>No</td><td>OAuth client ID.</td></tr><tr><td><code>client_secret</code></td><td>varchar(150)</td><td>No</td><td>OAuth client secret.</td></tr><tr><td><code>auth_username</code></td><td>varchar(150)</td><td>No</td><td>Authentication username.</td></tr><tr><td><code>auth_password</code></td><td>varchar(150)</td><td>No</td><td>Authentication password.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the MCP server was deleted.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who updated this object.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>Yes</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>Yes</td><td>Date this object was last updated.</td></tr></tbody></table>

### MCP Server Tool

The `mcp_server_tool` table stores a tool exposed by an MCP server.

Each tool includes the schema of the input it expects.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Tool UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this tool belongs to.</td></tr><tr><td><code>mcp_server_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the MCP server exposing the tool.</td></tr><tr><td><code>name</code></td><td>varchar(50)</td><td>Yes</td><td>Tool name.</td></tr><tr><td><code>description</code></td><td>text</td><td>Yes</td><td>Explains what the tool does.</td></tr><tr><td><code>input_schema</code></td><td>json</td><td>Yes</td><td>JSON schema describing the input expected by the tool.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who updated this object.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>Yes</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>Yes</td><td>Date this object was last updated.</td></tr></tbody></table>

## NLP Engine

The `nlp_engine` table stores cognitive engines configured for the environment.

A Conversational Project can use more than one engine. The record identifies the engine, its type, execution priority, and provider metadata.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>id</code></td><td>bigint unsigned</td><td>Yes</td><td>Cognitive engine ID.</td></tr><tr><td><code>engine</code></td><td>varchar(20)</td><td>Yes</td><td>Engine name, such as Syntphony NLP, Dialogflow, Watson, Luis, or KAI.</td></tr><tr><td><code>first</code></td><td>tinyint(1)</td><td>No</td><td>True when this engine is configured to run first.</td></tr><tr><td><code>type</code></td><td>varchar(10)</td><td>No</td><td>Engine type. Values can include <code>NLU</code>, <code>AGENTIC</code>, or <code>KAI</code>.</td></tr><tr><td><code>metadata</code></td><td>text</td><td>Yes</td><td>Provider-specific engine metadata.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who last updated this object.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr></tbody></table>

## Persona

The `persona` table stores the identity used by an agent.

A persona defines how an agent presents itself through its name, background, personality, communication style, and avatar. Every agent points to one persona.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Persona UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this persona belongs to.</td></tr><tr><td><code>name</code></td><td>varchar(50)</td><td>Yes</td><td>Persona name presented to the user.</td></tr><tr><td><code>backstory</code></td><td>text</td><td>No</td><td>Freeform background of the persona.</td></tr><tr><td><code>personality</code></td><td>text</td><td>No</td><td>Freeform description of the persona's personality.</td></tr><tr><td><code>communication_style</code></td><td>varchar(250)</td><td>No</td><td>Communication style, such as formal or casual.</td></tr><tr><td><code>avatar</code></td><td>varchar(255)</td><td>No</td><td>Path or URL of the persona image.</td></tr><tr><td><code>avatar_content_type</code></td><td>varchar(50)</td><td>No</td><td>Content type of the persona image.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>No</td><td>True when the persona was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

## Question

The `question` table stores questions associated with Knowledge AI.

A question points to a document whose content can be consulted when generating a response. Questions can also execute webhooks and store authentication information for those calls.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Question UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project.</td></tr><tr><td><code>document_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the document associated with the question.</td></tr><tr><td><code>name</code></td><td>varchar(255)</td><td>No</td><td>Question name.</td></tr><tr><td><code>description</code></td><td>varchar(255)</td><td>No</td><td>Question description.</td></tr><tr><td><code>webhook</code></td><td>varchar(255)</td><td>No</td><td>API URL called when the response is executed.</td></tr><tr><td><code>headers</code></td><td>longtext</td><td>No</td><td>Freeform webhook headers.</td></tr><tr><td><code>transactional</code></td><td>tinyint(1)</td><td>No</td><td>True when the response is transactional.</td></tr><tr><td><code>auth_type</code></td><td>enum</td><td>Yes</td><td>Authentication used when the question calls a webhook.</td></tr><tr><td><code>basic_token</code></td><td>varchar(2600)</td><td>No</td><td>Basic-auth token.</td></tr><tr><td><code>bearer_token</code></td><td>varchar(2600)</td><td>No</td><td>Bearer token.</td></tr><tr><td><code>o_auth_url</code></td><td>varchar(500)</td><td>No</td><td>OAuth token URL.</td></tr><tr><td><code>grant_type</code></td><td>enum</td><td>No</td><td>OAuth grant type.</td></tr><tr><td><code>client_id</code></td><td>varchar(150)</td><td>No</td><td>OAuth client ID.</td></tr><tr><td><code>client_secret</code></td><td>varchar(150)</td><td>No</td><td>OAuth client secret.</td></tr><tr><td><code>auth_username</code></td><td>varchar(150)</td><td>No</td><td>Authentication username.</td></tr><tr><td><code>auth_password</code></td><td>varchar(150)</td><td>No</td><td>Authentication password.</td></tr><tr><td><code>evaluable</code></td><td>tinyint(1)</td><td>No</td><td>True when the message can be evaluated.</td></tr><tr><td><code>enabled</code></td><td>tinyint(1)</td><td>No</td><td>True when the question is enabled.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>No</td><td>True when the question was deleted.</td></tr><tr><td><code>update_content</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the linked document changed and the question can be refreshed.</td></tr><tr><td><code>training_status</code></td><td>varchar(30)</td><td>No</td><td>Training status.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr></tbody></table>

### Question Variable

The `question_variable` table stores freeform variations associated with a question.

These variations represent different ways the same question can be expressed.

| Column Name         | Data Type    | Not Null | Description                                     |
| ------------------- | ------------ | -------- | ----------------------------------------------- |
| `uuid`              | varchar(36)  | Yes      | Variable UUID.                                  |
| `question_uuid`     | varchar(36)  | Yes      | UUID of the question this variation belongs to. |
| `question_variable` | varchar(255) | No       | Freeform question variation.                    |

## Satisfaction

The `satisfaction` table stores user feedback collected for a conversation.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Record UUID.</td></tr><tr><td><code>session_code</code></td><td>varchar(50)</td><td>Yes</td><td>Session code.</td></tr><tr><td><code>evaluation</code></td><td>tinyint(4)</td><td>Yes</td><td>User evaluation score.</td></tr><tr><td><code>answered</code></td><td>tinyint(1)</td><td>Yes</td><td>Indicates whether the user's question was answered.</td></tr><tr><td><code>user_comments</code></td><td>varchar(500)</td><td>No</td><td>User comments.</td></tr><tr><td><code>create_date</code></td><td>datetime</td><td>Yes</td><td>Creation date.</td></tr></tbody></table>

## Session

The `session` table stores one record for each conversation.

A session remains the common reference for all messages in that conversation and captures project, channel, user, locale, device, and runtime information.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>session_code</code></td><td>varchar(50)</td><td>Yes</td><td>Code representing the user session.</td></tr><tr><td><code>create_date</code></td><td>datetime</td><td>Yes</td><td>Session creation date.</td></tr><tr><td><code>last_message_date</code></td><td>datetime</td><td>No</td><td>Date of the latest interaction in the session.</td></tr><tr><td><code>expire_date</code></td><td>datetime</td><td>No</td><td>Session expiration date.</td></tr><tr><td><code>business_key</code></td><td>varchar(100)</td><td>No</td><td>Customer-defined identifier, such as a phone number or customer ID.</td></tr><tr><td><code>session_type</code></td><td>varchar(10)</td><td>Yes</td><td>Session state, such as created, updated, or expired.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project.</td></tr><tr><td><code>channel_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the channel.</td></tr><tr><td><code>operating_system</code></td><td>varchar(200)</td><td>No</td><td>Operating system.</td></tr><tr><td><code>operating_system_version</code></td><td>varchar(30)</td><td>No</td><td>Operating-system version.</td></tr><tr><td><code>browser</code></td><td>varchar(50)</td><td>No</td><td>Browser.</td></tr><tr><td><code>browser_version</code></td><td>varchar(30)</td><td>No</td><td>Browser version.</td></tr><tr><td><code>user_ref</code></td><td>varchar(50)</td><td>Yes</td><td>Channel-dependent user reference.</td></tr><tr><td><code>locale</code></td><td>varchar(6)</td><td>No</td><td>Language and country.</td></tr><tr><td><code>last_messages</code></td><td>text</td><td>No</td><td>Latest messages retained for the session.</td></tr><tr><td><code>test</code></td><td>tinyint(1)</td><td>Yes</td><td>True when the session comes from an automated test.</td></tr></tbody></table>

## Tags

The `tags` table stores tags created inside a Conversational Project.

Tags provide reusable labels that can be associated with supported objects.

| Column Name  | Data Type   | Not Null | Description                         |
| ------------ | ----------- | -------- | ----------------------------------- |
| `uuid`       | varchar(36) | Yes      | Tag UUID.                           |
| `bot_uuid`   | varchar(36) | Yes      | UUID of the Conversational Project. |
| `name`       | varchar(80) | Yes      | Tag name.                           |
| `created_at` | datetime    | No       | Date this object was created.       |
| `updated_at` | datetime    | No       | Date this object was last updated.  |

### Tag Use

The `tag_use` table stores the relationship between a tag and the object that uses it.

Each row points to one supported object.

Supported associations include:

* Intent
* Entity
* Answer
* Service
* REST connector
* Generative service
* Flow
* Document
* Question
* Supervisor
* Agent
* Agent function
* MCP server.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>id</code></td><td>bigint(20)</td><td>Yes</td><td>Relationship ID.</td></tr><tr><td><code>tag_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the tag.</td></tr><tr><td><code>tag_type_id</code></td><td>int</td><td>Yes</td><td>ID identifying the type of tagged object.</td></tr><tr><td><code>intent_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged intent.</td></tr><tr><td><code>entity_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged entity.</td></tr><tr><td><code>answer_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged answer.</td></tr><tr><td><code>service_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged service.</td></tr><tr><td><code>rest_connector_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged REST connector.</td></tr><tr><td><code>generative_service_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged generative service.</td></tr><tr><td><code>flow_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged flow.</td></tr><tr><td><code>document_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged document.</td></tr><tr><td><code>question_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged question.</td></tr><tr><td><code>supervisor_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged supervisor.</td></tr><tr><td><code>agent_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged agent.</td></tr><tr><td><code>function_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged agent function.</td></tr><tr><td><code>mcp_server_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the tagged MCP server.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr></tbody></table>

## Technical Log

The `technical_log` table stores technical information generated during execution, including response timing, service calls, errors, and troubleshooting identifiers.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>id</code></td><td>bigint(20)</td><td>Yes</td><td>Log ID.</td></tr><tr><td><code>session_code</code></td><td>varchar(50)</td><td>No</td><td>Session code when the log belongs to a conversation.</td></tr><tr><td><code>log_code</code></td><td>varchar(50)</td><td>Yes</td><td>Troubleshooting code associated with the log.</td></tr><tr><td><code>type</code></td><td>tinyint(4)</td><td>Yes</td><td>Log type: <code>0</code> error, <code>1</code> warning, <code>2</code> information.</td></tr><tr><td><code>message</code></td><td>text</td><td>No</td><td>Log message.</td></tr><tr><td><code>stacktrace</code></td><td>varchar(1000)</td><td>No</td><td>Stack trace when available.</td></tr><tr><td><code>module_type</code></td><td>varchar(50)</td><td>Yes</td><td>Architecture module that generated the log.</td></tr><tr><td><code>create_date</code></td><td>datetime</td><td>Yes</td><td>Creation date.</td></tr><tr><td><code>processing_time</code></td><td>int(11)</td><td>No</td><td>Processing time in milliseconds.</td></tr></tbody></table>

## Training

The `training` table stores training executions for Conversational Projects not using Knowledge AI.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Training UUID.</td></tr><tr><td><code>version</code></td><td>varchar(5)</td><td>Yes</td><td>Training version.</td></tr><tr><td><code>status</code></td><td>varchar(40)</td><td>Yes</td><td>Training status.</td></tr><tr><td><code>user_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who started the training.</td></tr><tr><td><code>accuracy</code></td><td>tinyint(4)</td><td>Yes</td><td>Training accuracy.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project.</td></tr><tr><td><code>updatedAt</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>error_msg</code></td><td>varchar(255)</td><td>No</td><td>Error message returned when training fails.</td></tr></tbody></table>

### Training AL

The `training_al` table stores Knowledge AI training executions.

A record identifies the Conversational Project, the collection being trained, the user who started the execution, and the training state for documents and FAQs.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Training UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project.</td></tr><tr><td><code>collection_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the collection being trained.</td></tr><tr><td><code>user_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the user who started the training.</td></tr><tr><td><code>version</code></td><td>varchar(5)</td><td>Yes</td><td>Training version.</td></tr><tr><td><code>status_document</code></td><td>varchar(40)</td><td>Yes</td><td>Document training status.</td></tr><tr><td><code>error_document</code></td><td>varchar(255)</td><td>No</td><td>Document training error.</td></tr><tr><td><code>status_faq</code></td><td>varchar(40)</td><td>Yes</td><td>FAQ training status.</td></tr><tr><td><code>error_faq</code></td><td>varchar(255)</td><td>No</td><td>FAQ training error.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr></tbody></table>

### Training AL Document

The `training_al_document` table stores the training state of each document participating in a Knowledge AI training execution.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Relationship UUID.</td></tr><tr><td><code>training_al_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Knowledge AI training.</td></tr><tr><td><code>document_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the document.</td></tr><tr><td><code>storage_filename</code></td><td>varchar(42)</td><td>No</td><td>Storage filename for the document version used during training.</td></tr><tr><td><code>document_training_status</code></td><td>varchar(30)</td><td>No</td><td>Training status for this document.</td></tr><tr><td><code>error_document</code></td><td>varchar(255)</td><td>No</td><td>Error returned when document training fails.</td></tr></tbody></table>

## Transactional Service

The `transactional_service` table stores external service calls configured in a Conversational Project.

The record defines the webhook, payload, headers, and authentication used by the service call.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Transactional service UUID.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project.</td></tr><tr><td><code>name</code></td><td>varchar(50)</td><td>Yes</td><td>Service-call name.</td></tr><tr><td><code>webhook</code></td><td>text</td><td>Yes</td><td>API URL to call.</td></tr><tr><td><code>content</code></td><td>text</td><td>Yes</td><td>Content sent during the service call.</td></tr><tr><td><code>headers</code></td><td>longtext</td><td>No</td><td>Headers required by the API call.</td></tr><tr><td><code>description</code></td><td>text</td><td>No</td><td>Service description.</td></tr><tr><td><code>auth_type</code></td><td>enum</td><td>Yes</td><td>Authentication method.</td></tr><tr><td><code>basic_token</code></td><td>varchar(2600)</td><td>No</td><td>Basic-auth token.</td></tr><tr><td><code>bearer_token</code></td><td>varchar(2600)</td><td>No</td><td>Bearer token.</td></tr><tr><td><code>o_auth_url</code></td><td>varchar(500)</td><td>No</td><td>OAuth token endpoint.</td></tr><tr><td><code>grant_type</code></td><td>enum</td><td>No</td><td>OAuth grant type.</td></tr><tr><td><code>client_id</code></td><td>varchar(150)</td><td>No</td><td>OAuth client ID.</td></tr><tr><td><code>client_secret</code></td><td>varchar(150)</td><td>No</td><td>OAuth client secret.</td></tr><tr><td><code>auth_username</code></td><td>varchar(150)</td><td>No</td><td>Authentication username.</td></tr><tr><td><code>auth_password</code></td><td>varchar(150)</td><td>No</td><td>Authentication password.</td></tr><tr><td><code>removed</code></td><td>tinyint(1)</td><td>No</td><td>True when the service was deleted.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

## User Interaction

The `user_interaction` table stores individual messages exchanged during a conversation.

Records that belong to the same conversation share the same `session_code`. In addition to message content, the table can identify the response, flow, Knowledge AI source, intent, entities, and confidence associated with the interaction.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Interaction UUID.</td></tr><tr><td><code>create_date</code></td><td>datetime</td><td>Yes</td><td>Date this interaction was created.</td></tr><tr><td><code>session_code</code></td><td>varchar(50)</td><td>Yes</td><td>Session code.</td></tr><tr><td><code>user_sent</code></td><td>tinyint(4)</td><td>Yes</td><td>True when the interaction is a user message.</td></tr><tr><td><code>answer_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the delivered answer.</td></tr><tr><td><code>answer_name</code></td><td>varchar(50)</td><td>No</td><td>Name of the delivered answer.</td></tr><tr><td><code>flow_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the flow that produced the interaction.</td></tr><tr><td><code>document_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the document used to build the response.</td></tr><tr><td><code>document_name</code></td><td>varchar(255)</td><td>No</td><td>Name of the document used to build the response.</td></tr><tr><td><code>document_paragraph</code></td><td>int(11)</td><td>No</td><td>Paragraph of the document used to build the response.</td></tr><tr><td><code>question_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the question associated with the interaction.</td></tr><tr><td><code>question_name</code></td><td>varchar(50)</td><td>No</td><td>Name of the associated question.</td></tr><tr><td><code>text</code></td><td>text</td><td>No</td><td>Conversation content.</td></tr><tr><td><code>button_texts</code></td><td>text</td><td>No</td><td>Text shown in interaction buttons.</td></tr><tr><td><code>intent</code></td><td>varchar(100)</td><td>No</td><td>Intent identified for the interaction.</td></tr><tr><td><code>highest_scored_intent</code></td><td>varchar(100)</td><td>No</td><td>Intent with the highest confidence.</td></tr><tr><td><code>confidence</code></td><td>decimal(12,10)</td><td>No</td><td>Confidence returned by the NLP.</td></tr><tr><td><code>highest_confidence_score</code></td><td>decimal(12,10)</td><td>No</td><td>Highest confidence returned by the NLP.</td></tr><tr><td><code>entities</code></td><td>text</td><td>No</td><td>Entities identified in the interaction.</td></tr><tr><td><code>user_sent_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user message that generated the response.</td></tr><tr><td><code>evaluation</code></td><td>tinyint(1)</td><td>No</td><td>Deprecated. Refer to <code>satisfaction</code>.</td></tr><tr><td><code>intent_uuid</code></td><td>varchar(36)</td><td>No</td><td>UUID of the intent identified in the interaction.</td></tr></tbody></table>

## Utterance

The `utterance` table stores example phrases associated with intents.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Utterance UUID.</td></tr><tr><td><code>intent_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the intent represented by this example.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project.</td></tr><tr><td><code>example</code></td><td>varchar(1000)</td><td>Yes</td><td>Example text.</td></tr><tr><td><code>metadata</code></td><td>varchar(100)</td><td>No</td><td>Metadata from imported NLP utterances.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

## Variable System

The `variable_system` table stores system variables available during a conversation.

These variables expose runtime information such as the session, channel, or locale through predefined keys.

| Column Name  | Data Type        | Not Null | Description                                                                   |
| ------------ | ---------------- | -------- | ----------------------------------------------------------------------------- |
| `id`         | tinyint unsigned | Yes      | System variable ID.                                                           |
| `name`       | varchar(50)      | Yes      | Variable name, such as `SessionCode` or `Locale`.                             |
| `key_name`   | varchar(50)      | Yes      | Key used to reference the variable, such as `$sessionCode` or `$info.locale`. |
| `created_at` | datetime         | Yes      | Date this object was created.                                                 |
| `updated_at` | datetime         | Yes      | Date this object was last updated.                                            |

## WhatsApp User

The `whatsapp_user` table stores user data associated with an Infobip WhatsApp channel.

| Column Name    | Data Type    | Not Null | Description            |
| -------------- | ------------ | -------- | ---------------------- |
| `phone number` | bigint(15)   | Yes      | User phone number.     |
| `create_date`  | datetime     | No       | Record creation date.  |
| `scenario_key` | varchar(100) | No       | WhatsApp scenario key. |

## Relationship tables

Relationship tables represent associations that cannot be expressed as a single owner field on one object.

### Agent–Agent Function relationship

The `agent_agent_function` table identifies which functions are available to each agent.

| Column Name           | Data Type   | Not Null | Description                                 |
| --------------------- | ----------- | -------- | ------------------------------------------- |
| `agent_uuid`          | varchar(36) | Yes      | UUID of the agent.                          |
| `agent_function_uuid` | varchar(36) | Yes      | UUID of the function assigned to the agent. |

### Agent–MCP Server Tool relationship

The `agent_mcp_server_tool` table identifies which MCP server tools an agent can call.

A tool cannot be assigned more than once to the same agent and MCP server.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>uuid</code></td><td>varchar(36)</td><td>Yes</td><td>Relationship UUID.</td></tr><tr><td><code>agent_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the agent that can call the tool.</td></tr><tr><td><code>mcp_server_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the MCP server exposing the tool.</td></tr><tr><td><code>mcp_server_tool_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the MCP server tool assigned to the agent.</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project this relationship belongs to.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>Yes</td><td>Date this object was created.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>Yes</td><td>Date this object was last updated.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr></tbody></table>

### Bot–NLP Engine relationship

The `bot_nlp_engine` table connects a Conversational Project with a cognitive engine.

A Conversational Project can use more than one engine.

<table data-header-hidden data-search="false"><thead><tr><th></th><th></th><th></th><th></th></tr></thead><tbody><tr><td>Column Name</td><td>Data Type</td><td>Not Null</td><td>Description</td></tr><tr><td><code>bot_uuid</code></td><td>varchar(36)</td><td>Yes</td><td>UUID of the Conversational Project.</td></tr><tr><td><code>nlp_engine_id</code></td><td>bigint unsigned</td><td>Yes</td><td>ID of the cognitive engine.</td></tr><tr><td><code>created_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who created this object.</td></tr><tr><td><code>updated_by</code></td><td>varchar(36)</td><td>No</td><td>UUID of the user who updated this object.</td></tr><tr><td><code>created_at</code></td><td>datetime</td><td>No</td><td>Date this object was created.</td></tr><tr><td><code>updated_at</code></td><td>datetime</td><td>No</td><td>Date this object was last updated.</td></tr></tbody></table>

### NLP Engine–Entity System relationship

The `nlp_engine_entity_system` table connects system entities to NLP engines.

When the relationship exists, the system entity is enabled for that NLP engine.

| Column Name        | Data Type           | Not Null | Description       |
| ------------------ | ------------------- | -------- | ----------------- |
| `nlp_engine_id`    | bigint(20) unsigned | Yes      | NLP engine ID.    |
| `entity_system_id` | bigint(20)          | Yes      | System entity ID. |


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs-conversational-ai.syntphony.com/api-docs/appendices/environment-data-structure.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
