Built by Metorial, the integration platform for agentic AI.
Retrieve all teams accessible with the current API key. Returns team details including name, status, plan info, and usage statistics (bot count, source count, question count).
Rate a bot's answer as positive, negative, or neutral (reset). Use the answerId from a chat response to submit the rating. Also supports recording a human support escalation for a given answer.
Add a new training source to a bot. Supports URL, sitemap, YouTube, RSS, Q&A pairs, and file-based sources. For file uploads, first use the **Get Upload URL** tool to obtain a presigned URL, upload the file, then pass the returned file path here.
Soft-delete a question from the bot's question history. The question is flagged as deleted and its content is cleared.
Retrieve full details of a conversation session including the complete message history, summary, sentiment, resolution/escalation status, and user metadata.
Permanently delete a bot and all its associated sources and data. This action cannot be undone.
List all training sources for a specific bot. Returns source type, status, title, URL, page/chunk counts, and schedule interval.
Update an existing bot's settings including name, description, AI model, custom prompt instructions, privacy, widget appearance (color, icon, alignment), allowed domains, branding, support link, and UI labels. Only provided fields are updated.
Delete a conversation session and all its messages from a bot. Requires bot edit permissions.
Delete a training source from a bot. The source data will be removed from the bot's index. Sources may still appear in chat results for several minutes while being fully removed.
Retrieve question and answer history for a bot. Supports filtering by rating, escalation status, answer capability, date range, and IP hash. Returns paginated results with question text, answer, sources, and metadata.
Retrieve full details for a specific bot including its configuration, model, status, widget settings, labels, and usage statistics.
Record that a chat interaction was escalated to human support. Call this when a user requests human assistance during a bot conversation.
Create a new AI chatbot for the configured team. Requires a name, description, privacy level, and language. Optionally specify the AI model or duplicate an existing bot.
Get a presigned upload URL for uploading a file to use as a bot training source. After uploading the file to the returned URL, use the returned file path with the **Create Source** tool.
Retrieve paginated conversation sessions for a bot. Returns conversation summaries including title, sentiment, resolution status, escalation status, and timestamps. Use **Get Conversation** to retrieve full message history.
List all bots for the configured team. Returns each bot's name, description, model, status, and usage statistics.
Send a question to a DocsBot AI agent and receive an answer with source citations. Supports multi-turn conversations by reusing the same conversationId, multimodal inputs via image URLs, and configurable parameters like reasoning effort and context items.