Three tools with the product name baked into them — agentic retrieval, full documents by URL, and a feedback channel.
LABSCO SUMMARY
Generic tool names collide the moment a client connects two documentation servers — two search_docs tools and the model picks wrong. Generating the product name into the tool name solves that at the source, which matters more as these become common. The per-call analytics redaction is the other thoughtful piece: a query can contain something confidential, and being able to suppress it per call rather than per account is the right granularity.
Kapa's own public documentation server. Customers get their own subdomain; this is the one Kapa runs for its own docs. The tool names are generated per product, so they read as search_<PRODUCT_NAME>_knowledge_sources (on Kapa's own docs server, search_kapa_knowledge_sources).
search_<PRODUCT_NAME>_knowledge_sources — agentic retrieval across every connected knowledge source, returning passages with their source URLsget_<PRODUCT_NAME>_knowledge_documents — the full document by its exact source URLOAuth 2.1 bearer on the docs server. Customer servers can instead use an API key or an internal Kapa login. With an API-key server, a per-call flag redacts the query text from Kapa's analytics.
Paste a URL, then authorize — add the endpoint to your client, then approve the OAuth consent screen
USE CASES
Retrieve passages with their source URLs, then fetch the full document if needed.
Product-named tools mean the agent picks the right one.
Set the redaction flag on that call.
Listed in project docs. Your client may show a different list.
search_<PRODUCT_NAME>_knowledge_sourcesAgentic retrieval over all knowledge sources connected to the Kapa project. Returns most relevant chunks in descending relevance; each chunk is a short self-contained snippet from a single page/item. Results are a structured list of {source_url, content (Markdown)}. Wraps the Retrieval API; p50 ~3s, p95 ~4.5s latency; use_pruning adds ~0.7s.
get_<PRODUCT_NAME>_knowledge_documentsFetch full documents from the knowledge sources by their exact source URL. Returns full content in Markdown; non-exact URL matches are omitted (results may be empty). Paginates and truncates long documents so the agent context is not flooded. Returns {source_url, title, content}.
give_feedbackLets the calling agent file actionable feedback on the product, the documentation, or the MCP server itself against the Kapa project. Does not retrieve knowledge; the agent gets a short confirmation.
BEFORE YOU INSTALL
3 tools named in the project's documentation.
INSTALLATION
Add the endpoint to your client, then approve the OAuth consent screen.
https://kapa-docs.mcp.kapa.aiMCP servers
kubernetools
lhstorm
TechDocsStudio