Integration

Chat Data integration

Chat Data builds AI chatbots trained on your own documents, websites, products and Q&As. Create, train, retrain and configure chatbots, ask them questions, read conversations, leads (customers) and analytics, take over a conversation in live chat, pause the bot, send outbound WhatsApp Web messages, manage AI actions, webhooks, uploaded files and white-label sub-users. Profile picture and chat icon uploads are not exposed, and the reply to a chat is returned in its OpenAI-compatible JSON form (streaming is not supported). Authentication is an API key sent as a Bearer token: a workspace key reaches every chatbot, a chatbot-scoped key only its own chatbot.

Connect Chat Data How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Chat Data.Built from the 39 Chat Data tools below, on the connection you add once.

What you can do with Chat Data

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Sub user, 4 operations

create_sub_userCreate sub userPOSTCreates a sub-user who can log in to your branded portal and use the assigned chatbots; without a password they sign in by magic link sent to their email. Needs the SSO Login add-on and a workspace API key.
delete_sub_userDelete sub userDELETEDeletes a sub-user and revokes their pending magic-link and password-setup tokens. There is no tool to restore it. Needs the SSO Login add-on and a workspace API key.
list_sub_usersList sub usersGETLists the sub-users of the workspace's company (white-label portal logins) with their emails, assigned chatbots and permissions; empty when the workspace has no SSO company. Needs the SSO Login add-on and a workspace key.
update_sub_userUpdate sub userPUTUpdates a sub-user's name, password, chatbots, permissions or analytics sections. Only the fields you send change; the list fields (chatbotIds, permissions, analyticsDashboardSections) replace the current list. Needs SSO Login.

Get chatbot, 2 operations

get_chatbotGet chatbotGETReturns one chatbot with its settings, data sources and training status (status 100 means training has finished).
list_chatbotsList chatbotsGETLists the chatbots the API key can reach (every chatbot of the workspace for a workspace key, only its own for a chatbot-scoped key) with their settings, data sources and training status.

Get conversation, 2 operations

get_conversationGet conversationGETReturns one conversation with its messages, lead details and metadata. Without chatbotId it is looked up across every chatbot the API key can reach. The lead details can hold personal data (name, email, phone).
list_conversationsList conversationsGETLists a chatbot's conversations, most recently active first, with their messages. Only conversations with at least one user message are returned. Paginate with start and size; total is the full count.

Add action, 1 operation

add_actionAdd actionPOSTAdds an AI action (api, search, lead or button) to a chatbot. An api action is test-called once before it is saved: Chat Data sends a real request to its url, so point it only at an endpoint that may receive one.

Add webhook, 1 operation

add_webhookAdd webhookPOSTSubscribes a URL to one chatbot event (chat, live-chat-escalation, lead-submission, user-online, user-offline); call again to add another event to the same url. Needs a plan with webhooks; at most 5 endpoints per chatbot.

Attach lead, 1 operation

attach_leadAttach leadPOSTAttaches a lead (name, email, phone and other identifiers) to a conversation, or updates the lead it already has: the fields you send are merged into it. A lead's uuid can never be changed once set.

Bot control, 1 operation

set_bot_controlSet bot controlPOSTPauses (-1) or resumes (1) the bot in one conversation from WhatsApp, Instagram, Messenger, WhatsApp Web API, Telegram or LINE; omit botControl to toggle. While paused the bot does not answer user messages.

Create chatbot, 1 operation

create_chatbotCreate chatbotPOSTCreates a chatbot and starts training it. model is required: custom-data-upload (your text, URLs, products, Q&As), medical-chat-human, medical-chat-vet or custom-model (your own backend). Training consumes the plan's training characters.

Current plan, 1 operation

get_current_planGet current planGETReturns the workspace's subscriptions, remaining message credits and training characters, and plan limits and features, plus the owner's name and email. Needs a workspace API key (403 for a chatbot-scoped key).

Delete action, 1 operation

delete_actionDelete actionDELETEDeletes one AI action of a chatbot, chosen by name or by index (send exactly one). There is no tool to restore it.

Delete chatbot, 1 operation

delete_chatbotDelete chatbotDELETEPermanently deletes a chatbot with all its data: training data, conversations, leads and analytics. It cannot be recovered.

Delete conversation, 1 operation

delete_conversationDelete conversationDELETEDeletes one conversation of a chatbot. Both chatbotId and conversationId are required. There is no tool to restore it.

Delete file, 1 operation

delete_fileDelete filePOSTDeletes a file previously uploaded with upload_file, by its url. There is no tool to restore it.

Delete webhook, 1 operation

delete_webhookDelete webhookPOSTRemoves the webhook registered with the given url, with all its event subscriptions, and returns the remaining webhooks. Permanent. If no webhook matches, nothing is deleted and the call still succeeds.

Get action, 1 operation

get_actionGet actionGETReturns one AI action of a chatbot, chosen by name or by index (send exactly one of them).

Get admin live chat link, 1 operation

get_admin_live_chat_linkGet admin live chat linkGETReturns a link to the admin live chat portal for one conversation, where an agent can answer the user. Needs a plan with live chat and live chat enabled in the chatbot's settings. chatbotId is required.

Get analysis, 1 operation

get_analysisGet analysisGETReturns a chatbot's aggregated statistics: messages, credits, conversations, leads, live chat escalations, votes, topics, sentiments, ratings and countries, per day or per hour. Needs a plan with user analysis.

Get confidence score, 1 operation

get_confidence_scoresGet confidence scoresPOSTScores how well a chatbot's knowledge base matches the last message, without generating an answer: returns per-chunk scores, the average and the matched sources. Use it before send_message to judge answer quality.

Get conversation portal link, 1 operation

get_conversation_portal_linkGet conversation portal linkGETReturns a shareable link, valid 30 days, to the conversation portal for one conversation (the portal must be enabled for the chatbot). The conversation is not looked up, so a link is returned even for an unknown id.

Get customer, 1 operation

list_customersList customersGETLists the customers (leads) a chatbot collected, newest first, with filters on source, lead id, name, email, phone and usernames. Returns personal data (names, emails, phone numbers). Paginate with start and size.

Get sas token, 1 operation

get_file_access_urlGet file access urlPOSTReturns a read-only URL with a fresh access token for one uploaded file (file urls in conversations are private). Anyone holding the returned url can read the file. For a missing file the token part is empty.

List action, 1 operation

list_actionsList actionsGETLists the AI actions configured on a chatbot (api, search, lead and button actions), with their parameters.

Live chat, 1 operation

send_live_chat_messageSend live chat messagePOSTSends a live chat message into an existing conversation, as a human agent (default; the bot then stops answering), the assistant or the user. On WhatsApp, Messenger, Instagram, Telegram or LINE conversations it is delivered to the user.

Livechat state, 1 operation

set_live_chat_stateSet live chat statePOSTEnds (end true) or reopens (end false) the live chat session of a conversation; omit end to toggle. Needs a plan with live chat and live chat enabled in the chatbot's settings.

Message, 1 operation

send_messageSend messagePOSTAsks a chatbot a question; the answer comes in OpenAI's chat completion format, reply text in choices[0].message.content (the vendor prints no example). Uses message credits. Pass conversationId to keep one conversation.

Query knowledge base, 1 operation

query_knowledge_baseQuery knowledge basePOSTSearches a chatbot's knowledge base by vector similarity and returns the matching segments (source, type, snippet, score) without generating an answer. Results are cached by query.

Retrain chatbot, 1 operation

retrain_chatbotRetrain chatbotPOSTAdds to or removes from a custom-data-upload chatbot's knowledge base (text, URLs, products, Q&As; deletes removes sources) and retrains it. Training consumes characters; poll get_chatbot_status for progress.

Send whatsapp web message outbound, 1 operation

send_whatsapp_web_messageSend whatsapp web messagePOSTSends a real WhatsApp Web message (text, or audio by text-to-speech) to a phone number from a chatbot's connected WhatsApp Web account, creating the conversation and lead if needed. Returns conversationId and leadUUID.

Status, 1 operation

get_chatbot_statusGet chatbot statusGETReturns a chatbot's training progress, from 0 to 100; 100 means no training is running. Poll it after create_chatbot or retrain_chatbot.

Toggle bot activation, 1 operation

set_integration_activationSet integration activationPOSTTurns a chatbot's WhatsApp, Messenger, Instagram or LINE integration on or off (active true or false; omit to toggle). Telegram and WhatsappWebAPI cannot be toggled: the call returns active null and changes nothing.

Update chatbot settings, 1 operation

update_chatbot_settingsUpdate chatbot settingsPOSTChanges a chatbot's settings (name, prompt, model, messages, domains, rate limits, widget style, lead form, booking buttons). Only the fields you send change; list fields replace the current list; calCom is replaced as a whole.

Update metrics share, 1 operation

update_metrics_shareUpdate metrics sharePOSTMakes a chatbot's analytics dashboard public or private and picks the graphs it shows; always returns the public dashboard link. Making it public needs a plan with analytics sharing. Only the fields you send change.

Update white label, 1 operation

update_white_labelUpdate white labelPOSTChanges a chatbot's white-label settings: Powered-by text, company name and URL, custom embedding domain and sender email (the response lists the DNS records to add). Only sent fields change; each needs the matching plan add-on.

Upload file, 1 operation

upload_fileUpload filePOSTUploads a file into a chatbot's folder and returns its url, to attach as files in send_message or send_live_chat_message. A file with the same name is not replaced: the existing file's url is returned instead.

Use Chat Data from Claude Code, Cursor or Codex

Connect Chat Data in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Chat Data operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect Chat Data

Create your workspace
Sign up free. Nothing to install.
Connect Chat Data once
Add your Chat Data connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.