Hey makers and developers!
Over the past year, nearly every documentation platform rolled out an AI chatbot. But almost all of them followed the exact same monetisation playbook: high monthly subscriptions ($100 to $500/mo), query caps, and vendor lock-in on your own search index.
With docmd, we took the opposite approach: what if the AI assistant was simply a native, free feature of the open-source documentation generator itself?
Hey Product Hunt! 👋
Ghazi here from docmd.
When we built docmd, our goal was simple: eliminate the boilerplate, framework bloat, and configuration overhead from documentation.
Today, we are taking the next big step by launching docmd Assistant and making docmd completely AI-native.
Documentation is no longer just read by developers in browsers, it is queried by IDE copilots, indexed by AI agents, and referenced in chat workflows. But adding AI to docs has traditionally meant paying $200+ a month to a third-party SaaS and giving up data privacy.
With docmd Assistant, we built AI directly into the open-source engine:
Bring Your Own Key (BYOK): Connect OpenAI, Anthropic Claude, Google Gemini, DeepSeek, Groq, or local Ollama. No monthly subscriptions, no query markups, and complete data privacy.
Native MCP Server (`docmd mcp`): Lets Claude Desktop, Cursor, and AI agents query your documentation directly through standard Model Context Protocol.
Automated AI Manifests: Auto-generates `llms.txt` and OKF bundles on every build.
Embeddable Anywhere: Works natively in docmd or embeds as a lightweight widget on any existing site.
Everything is 100% free and open-source under the MIT license.
Would love to hear your thoughts and feedback on how you are using AI in your docs!
Hi! This looks super interesting! I am currently using Astro, but this might actually be a reason to spend the 20 minutes migrating.
@foobar_beer You'll absolutely love the experience, active community that listens and actually builds on requests with regular updates. Cheers!!