Documentation built for humans, AI assistants, and agents
dubstack turns developer documentation into a reliable source of truth that can be read by people, searched by assistants, and queried by agents through public MCP endpoints.
Key points
Publish AI-ready docs from Git-backed MDX and structured configuration.
Give users an assistant that answers from the published docs and links back to source pages.
Expose public read-only MCP endpoints so coding agents can search and read documentation directly.
Use workflows to keep docs fresh as code, feedback, and product changes land.
Why AI-ready docs matter
AI tools are becoming a primary interface for developer research. If documentation is not structured, crawlable, and source-linked, assistants either ignore it or answer from stale third-party context.
- -Clear canonical pages help AI systems identify the product and its use cases.
- -Source-linked assistant responses build trust with developers and support teams.
- -Agent-readable documentation reduces repeated manual answers from engineering teams.
How dubstack supports AI experiences
dubstack combines hosted documentation, search, assistant retrieval, MCP, and Git workflows so the same source content can support multiple AI surfaces.
- -Published docs are available to the user-facing assistant.
- -Public MCP endpoints let external agents search and read the docs site.
- -Dashboard workflows can draft documentation updates for review instead of silently changing production docs.
Questions
What makes dubstack AI-ready?
dubstack focuses on structured docs, source-linked retrieval, public MCP access, and Git-backed workflows so AI systems can read and cite authoritative documentation.
Does dubstack replace human documentation review?
No. AI workflows draft and assist, while Git-backed review keeps humans in control of production documentation changes.