Chat Data

Send a message to a Chat Data chatbot and receive the AI-generated response, with optional streaming and conversation continuity.

Retrieve a paginated list of all chatbot conversations, including metadata like timestamps, lead info, and escalation status.

Fetch the full transcript and metadata for one specific conversation by ID, including all messages and any attached lead data.

Add a new message to an existing conversation thread, preserving context and continuing the AI dialogue without starting a new session.

Provision a new Chat Data chatbot with a specified knowledge source, model configuration, and initial prompt — ready to deploy in minutes.

Update a chatbot's knowledge base with new source text, URLs, or Q&A pairs — with the option to delete specific outdated chunks.

Modify a chatbot's name, base prompt, visibility, allowed domains, lead capture config, appearance styles, or notification settings.

Replace or refine a chatbot's system-level instruction prompt, instantly changing its persona, tone, constraints, and response boundaries.

Check the current training progress of a chatbot after a retrain is triggered, returning status and completion percentage.

List all chatbots in your Chat Data workspace with their names, IDs, statuses, and configuration metadata.

Retrieve full configuration details for a specific chatbot, including its base prompt, model settings, data sources, and embed configuration.

Permanently remove a chatbot and all its associated data — use for decommissioning deprecated bots after a grace period.

Retrieve all leads captured by your chatbots, including name, email, and the conversation context in which they were submitted.

Perform a semantic search against a chatbot's knowledge base and retrieve the most relevant context chunks for a given query.

Calculate cosine similarity scores between a query and your chatbot's knowledge base chunks before sending, revealing how well-covered a question is.

Upload a PDF, text, or other supported file to Chat Data so it can be ingested as a training source for a chatbot.

Send a message from the human-agent side of a live chat session, injecting your responses into an active escalated conversation.

Switch a conversation between AI-handled and human-live-chat mode, enabling programmatic handoffs in either direction.

Register a webhook endpoint on a chatbot to receive real-time event notifications for new messages, lead submissions, or live-chat escalations.

Retrieve performance metrics for a chatbot, including message volume, topic distribution, confidence score distribution, and user engagement stats.

Fires when a visitor sends a message to any of your Chat Data chatbots; can be scoped to a specific chatbot by ID.

Fires when a visitor completes the lead-capture form embedded in a Chat Data chatbot conversation.

Fires when the chatbot determines it cannot resolve a visitor's query and requests a human agent takeover.

Fires after Chat Data finishes ingesting new training data and a chatbot's model is updated.

Fires when the chatbot scores its own reply below the confidence threshold you configure, signalling a knowledge gap.

Fires when a chatbot is toggled off or deleted in the Chat Data dashboard.

Fires when a new webhook endpoint is successfully added to a Chat Data chatbot.