About Kapa
Kapa combines a knowledge ingestion and retrieval platform with an MCP server and agent skills. Teams can configure sources, search indexed content, and inspect analytics through an agent using their Kapa permissions.
Key features
- 01Index and update knowledge sources
- 02Configure sources in a Kapa project
- 03Search indexed content
- 04Review question and coverage analytics
- 05Use agent skills for platform workflows
Use cases
- 01Add documentation or support sources to a project
- 02Find answers across indexed knowledge
- 03Identify common questions and coverage gaps
MCP support
The Kapa platform MCP server exposes source configuration, knowledge search, and analytics capabilities to MCP clients. Agents can use it with the user’s Kapa permissions.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://mcp.kapa.ai/mcpclaude mcp add --transport http kapa https://mcp.kapa.ai/mcp{"mcpServers":{"kapa":{"type":"http","url":"https://mcp.kapa.ai/mcp"}}}Open in Cursor{"servers":{"kapa":{"type":"http","url":"https://mcp.kapa.ai/mcp"}}}Open in VS Code{"mcpServers":{"kapa":{"type":"http","url":"https://mcp.kapa.ai/mcp"}}}{"mcpServers":{"kapa":{"serverUrl":"https://mcp.kapa.ai/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEKapa plugin
A plugin that bundles the Kapa platform MCP server with skills for agents, so your agent can set up knowledge sources, search them and read analytics in a Kapa project. Sign in once and the tools act as you, with your Kapa permissions.
Kapa is an ingestion and retrieval system. Ingestion builds one knowledge base from 20+ types of sources and keeps it up to date as that content changes. Agentic retrieval searches across it to find what an agent needs. You hand Kapa to your agents as a search tool via MCP or the API, or use one of the prebuilt agents, such as a website widget or a Slack bot.
Install
Claude Code
/plugin marketplace add kapa-ai/kapa-plugin
/plugin install kapa
Codex
codex plugin marketplace add kapa-ai/kapa-plugin
codex plugin add kapa@kapa
Cursor
Inside Cursor's agent chat:
/add-plugin kapa
Or install the plugin from its Cursor Marketplace listing.
Every other client
Add the MCP server to your client's configuration:
{ "mcpServers": { "kapa": { "url": "https://mcp.kapa.ai/mcp" } } }
Then copy skills/ into wherever that client reads skills from, so the agent
gets the call sequences rather than only the tool list.
What it can do
The plugin bundles the Kapa platform MCP server with skills for working with the platform, and lets you do the following.
Configure sources
Index your knowledge sources in a Kapa project. You can create new sources or edit the configuration of existing ones.
Start with kapa-setup, or ask: "add our documentation site to Kapa". The
agent asks what you want to ingest, then loads the setup skill for that
source.
The following knowledge sources are supported:
- Web crawl
- GitHub files
- GitHub issues
- GitHub discussions
- GitHub pull requests
- Zendesk Help Center
- Zendesk tickets
- Confluence
- Jira
- Jira Service Management
- Notion
- Slack
- Discord
- Discourse
- Google Drive
- OpenAPI
- YouTube
- S3
- Hand-written answers
Search your knowledge base
All indexed sources can be searched with agentic retrieval through this plugin. For example, ask: "what does our knowledge base say about SSO?"
Searching through the plugin is mainly for testing what your knowledge base returns. To put it in front of users or agents, use the prebuilt agents, a hosted MCP server or the retrieval API.
Read analytics
Every question that flows through Kapa, whichever agent or interface asked it,
is kept as a conversation. This plugin gives your agent access to Kapa's
analytics, so it can pull the
top questions and
coverage gaps for a period, or
look through conversations.
kapa-gaps combines the latest gaps and top questions into suggestions for
what to write next.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.