Integration

kapa.ai integration

kapa.ai turns a company's technical knowledge (documentation, code, tickets, forums, PDFs and other sources) into a knowledge base and answers questions from it with citations. Ask the project's assistant a question or a follow-up in a thread, search the knowledge base for relevant chunks or fetch whole indexed documents, record upvote or downvote feedback, and read the project's analytics: threads, end users, activity statistics, top questions and coverage gaps, plus its sources, source groups and integrations. The streamed chat routes and the routes kapa.ai has deprecated (keyword search, the old activity and question-answer lists) are not exposed. Authentication is an API key sent in the X-API-KEY header, created in the Kapa platform under Deploy, API Keys; the project id is under Settings, Projects.

Connect kapa.ai How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in kapa.ai.Built from the 17 kapa.ai tools below, on the connection you add once.

What you can do with kapa.ai

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.

Each badge shows the default agent permission of an endpoint (allowed: an agent calls it without asking; ask first: you are asked before an agent calls it, when the agent or chat asks permission for sensitive actions; never: no agent may call it), which every user can change endpoint by endpoint in their credential settings.

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

Coverage gaps period, 2 operations

get_coverage_gaps_periodGet coverage gaps periodGETAgents: ask firstReturns the coverage-gap clusters of one period (questions answered with uncertainty): title, summary, a suggestion for closing the gap, thread and user counts and up to threads_limit recent threads. Period ids come from list_coverage_gaps_periods.
list_coverage_gaps_periodsList coverage gaps periodsGETAgents: ask firstLists completed coverage-gap periods of a project (weekly, monthly or quarterly clusterings of the questions answered with uncertainty), newest first. Pass a period id to get_coverage_gaps_period.

Question, 2 operations

ask_follow_up_questionAsk follow up questionPOSTAgents: ask firstAsks a follow-up question in an existing thread (thread_id from ask_question or list_threads) and returns the answer with its relevant sources, continuing that conversation.
ask_questionAsk questionPOSTAgents: ask firstAsks the project's kapa.ai assistant a question and returns the generated answer with its relevant sources. Starts a new thread, recorded in the project's analytics; pass the returned thread_id to ask_follow_up_question.

Thread, 2 operations

get_threadGet threadGETAgents: ask firstReturns one conversation thread: its integration and every question-answer pair with feedback and relevant sources. Set include_source_content to false to leave out each source's full text, which can be very long.
list_threadsList threadsGETAgents: ask firstLists the project's conversation threads with their question-answer pairs, most recent activity first; filter by integration, tags or updated_since and add nested data with include. Can hold end users' personal data.

Document, 1 operation

get_documentsGet documentsPOSTAgents: ask firstReturns the full Markdown of indexed documents, by URL (as search_project_knowledge returns them) or by document id, up to 5 per page; each is cut to max_chars_per_document and flagged truncated. Unmatched requests are left out.

End user, 1 operation

list_end_usersList end usersGETAgents: ask firstLists the end users who asked the project's assistant questions: email and unique_client_id when tracked, company name, question count, first and latest activity, latest thread and its first question. Personal data.

Feedback, 1 operation

upsert_feedbackUpsert feedbackPOSTAgents: ask firstRecords an end user's upvote or downvote, with an optional comment, on one answer (question_answer_id from ask_question); calling again for the same user_identifier and answer replaces that feedback. Shows in the project's analytics.

Integration, 1 operation

list_integrationsList integrationsGETAgents: ask firstLists the project's integrations (website widget, Slack, Discord, API, MCP server, Zendesk agent and others): id, name, type and dates. Pass an id as integration_id so chat and retrieval traffic is attributed to it.

Knowledge, 1 operation

search_project_knowledgeSearch project knowledgePOSTAgents: ask firstSearches the project's knowledge base and returns the most relevant chunks (source_url and Markdown content) with no generated answer, for use as context. Each call counts as a query in the project's analytics.

Project, 1 operation

get_projectGet projectGETAgents: ask firstReturns one kapa.ai project the API key can access: its name, team id, whether it is labeled internal, and its PII masking settings (entity types redacted, custom regexes, allow list).

Project activity, 1 operation

get_project_activityGet project activityGETAgents: ask firstReturns aggregate activity of a project over an optional date range: conversation and retrieval query counts, upvotes, downvotes, uncertain answers, unique users, support tickets deflected and queries by language, overall and per integration.

Source, 1 operation

list_sourcesList sourcesGETAgents: ask firstLists the project's knowledge sources (web crawl, GitHub, Slack, Zendesk, Confluence, Notion, file upload and more): id, name, type, whether marked internal, and the source groups each belongs to. Ids filter retrieval via source_ids_include.

Source group, 1 operation

list_source_groupsList source groupsGETAgents: ask firstLists the project's source groups (product or version groupings of its knowledge sources) with nested sub-groups. Pass group ids in source_group_ids_include of search_project_knowledge, get_documents or ask_question to limit them.

Use kapa.ai from Claude Code, Cursor or Codex

Connect kapa.ai in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its kapa.ai 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 kapa.ai

Create your workspace
Sign up free. Nothing to install.
Connect kapa.ai once
Add your kapa.ai 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.