Integration · Customer support

Connect Claude & ChatGPT to LiveChatAI

Ask your AI how your LiveChatAI agents are doing, read a conversation, or reply to a customer as a team member. It can filter conversations, manage contacts, add Q&As, texts and web pages to an agent's training data, and pull analytics on topics, sentiment, CSAT and resolutions.

Connect LiveChatAI free

Free plan, no credit card. Takes about a minute.

Try asking

“Show me open conversations from today with negative sentiment.”
The AI lists the agent's conversations filtered by status, date and sentiment, sorted by last message.
“Reply to that customer that the refund is on its way.”
It sends the reply as the team member you name. The customer sees it immediately.
“Teach our support agent that we are closed on 24 December.”
It adds a question and answer pair to the agent's training data, which the agent uses in future replies.

What your AI can do

Conversations and contacts

List and read conversations, reply as a team member, close conversations, and create, update and look up contacts.

Training data

Search an agent's data sources, add or update Q&As and texts, add web pages, and read AI-suggested improvements.

Analytics

Get overview analytics, CSAT ratings, topics, sentiment and AI versus human resolutions for any date range.

“Train the agent on our new returns page.”
It adds the page URL to the agent's training data so it gets crawled.
“How many conversations did the AI resolve last month, and how many went to a human?”
It reads the resolution analytics for the agent in that date range.
“What did customers ask about most this week?”
It gets the topic analytics for the agent between the dates you give.

25 tools for LiveChatAI

These tools are switched on when you connect. You can switch any of them off, or require your approval before it runs.

add_qaChanges dataTeach an agent a question and answer pair: `agent_id`, `question` and `answer`. The agent uses it in future replies.
add_textChanges dataAdd a plain-text training entry to an agent: `agent_id`, `title` and `content`.
add_website_urlChanges dataTrain an agent on a web page: `agent_id` and the page `url` to crawl.
ask_agentChanges dataSend a question to an AI agent and get its answer: `agent_id` and `question`, optionally continuing an existing `conversation_id`. Use to test how the agent answers; it is logged as a conversation.
create_or_update_contactChanges dataCreate a contact for an agent, or update one by passing its `contact_id`: `agent_id`, your own unique `distinct_id` for the person, plus name, email, phone and custom attributes.
end_conversationChanges dataClose a conversation by `conversation_id`, optionally recording which team member (`user_id`) closed it.
get_agentGet one AI agent's details and settings by `agent_id`.
get_analyticsGet an agent's overview analytics (conversations, messages, handled by AI vs humans) for `start_date` to `end_date`.
get_contactGet one contact by `contact_id`, with name, email, phone and custom attributes.
get_conversationGet one conversation with its messages, by conversation ID in `message_id`.
get_resolution_analyticsGet how many conversations the AI resolved vs handed to a human, for an agent between `start_date` and `end_date`.
get_sentiment_analyticsGet conversation sentiment (positive, negative, neutral) for an agent between `start_date` and `end_date`.
get_topic_analyticsGet the topics customers talk about with an agent between `start_date` and `end_date`.
list_agentsList your AI agents (chatbots) with their IDs. Start here: most other tools need an `agent_id`.
list_ai_suggestionsList AI-suggested improvements to an agent's answers by `agent_id`, optionally within a date range.
list_contact_conversationsList all conversations with one contact, by `contact_id`, with optional status, channel, sentiment and date filters.
list_contactsList an agent's contacts (chat visitors) by `agent_id`, with `page` and `limit`.
list_conversationsList an agent's conversations by `agent_id`, filtered by status (open/close), channel, sentiment, topic, date range, who handled it (AI, human) or resolution. Sorted by last message.
list_csat_ratingsList customer satisfaction (CSAT) ratings for an agent between `start_date` and `end_date`.
list_qasList an agent's Q&A training entries by `agent_id`.
list_textsList an agent's plain-text training entries by `agent_id`.
reply_to_conversationChanges dataSend a reply as a human team member in a conversation: `conversation_id`, `user_id` of the team member, and `message`. The customer sees it immediately.
search_training_dataSearch an agent's training data sources (websites, files, texts, Q&As) by `agent_id`, with optional `search` text and `source` type.
update_qaVisible to othersChange an existing Q&A training entry: `agent_id`, entry `id`, new `question` and `answer`.
update_textVisible to othersReplace a text training entry's title and content: `agent_id`, entry `id`, `title` and `content`.

Set up in three steps

1
Pick the app
Create a free PipMCP account and choose LiveChatAI from the app list.
2
Paste your key
Paste a LiveChatAI API key. In LiveChatAI open Dashboard › Profile › Personal and create or copy an API key.
3
Add the link to your AI
You get a personal MCP link. Add it to Claude, ChatGPT or Cursor:
1. Click your name, then Settings › Connectors › Add custom connector.
2. Paste your link as the Remote MCP server URL.
3. Switch it on from the + menu in a chat.

Questions

What can the AI do in LiveChatAI?

The AI can read your agents, conversations, contacts, training data, AI suggestions and analytics. It can reply in a conversation as a team member, close conversations, create and update contacts, add and update Q&As and texts, add web pages to training, and ask an agent a question.

Does the AI see my LiveChatAI API key?

No. Your LiveChatAI API key is encrypted at rest and never shown to the AI. After you save it, it is not shown again, not even to you. The AI only sees the results of the tools it calls.

Can I control what the AI is allowed to do?

Yes. You choose which tools are switched on, so you can start with reading conversations and analytics. Replies reach customers right away and training changes what your agent says, so those tools can require your approval before they run. Every tool call is logged.

Does it work with ChatGPT?

Yes. In ChatGPT go to Settings › Apps & Connectors › Advanced and turn on Developer mode, then add your PipMCP link. Developer mode needs a paid ChatGPT plan: Plus, Pro, Business or Enterprise. The same link also works in Claude (Settings › Connectors › Add custom connector), Cursor and other MCP clients.

What does it cost?

PipMCP has a free plan with no credit card required. Paid plans bill per completed task. You also need your own LiveChatAI account.

Can I test how my agent answers?

Yes. The ask_agent tool sends a question to one of your LiveChatAI agents and returns its answer, and it can continue an existing conversation. The question is logged as a conversation in LiveChatAI, so you can see it there too.