Connect AI tools like Cursor, VS Code, and Claude Desktop to your product docs via a hosted or self-hosted Retrieval-Augmented Generation (RAG) MCP server.
https://github.com/TechDocsStudio/biel-mcpYour AI assistant gives you generic answers because it doesn't know your product. You're constantly alt-tabbing between your IDE and docs, breaking flow every time you need to check an API endpoint or integration pattern.
The Biel MCP Server connects your AI coding tools directly to your product documentation, turning generic assistants into product-aware coding partners.
Instead of generating boilerplate code that might work, your AI assistant now references your actual API specs, authentication patterns, and integration examples. Ask "What are the required headers for the /users endpoint?" and get your exact API documentation, not a best guess.
This isn't about adding another tool to your stack—it's about making the tools you already use significantly more useful.
Skip the documentation hunt: Your AI assistant knows your product's authentication flow, error codes, and edge cases. No more digging through docs to understand why your integration isn't working.
Accurate code completions: When Cursor or VS Code suggests code, it's based on your actual API patterns and naming conventions, not generic examples from training data.
Onboarding acceleration: New team members get context-aware help instead of generic Stack Overflow answers. The AI can walk them through your specific setup patterns and configuration requirements.
Drop a single JSON configuration into Claude Desktop, Cursor, or VS Code. Your AI assistant immediately gains access to your documentation without changing your existing workflow.
{
"mcpServers": {
"biel-ai": {
"description": "Query your product's documentation, APIs, and knowledge base.",
"command": "npx",
"args": [
"mcp-remote",
"https://mcp.biel.ai/sse?project_slug=YOUR_PROJECT_SLUG&domain=https://your-docs-domain.com"
]
}
}
}
Then ask questions like:
Your AI assistant pulls answers directly from your documentation instead of hallucinating.
API Integration: Building against your own APIs becomes straightforward when your AI knows the exact request/response formats, required headers, and error handling patterns.
Configuration Management: Instead of memorizing environment variables and config options, ask your AI assistant about deployment settings, feature flags, or service configurations.
Troubleshooting: When something breaks, your AI can reference your troubleshooting guides, known issues, and resolution patterns instead of suggesting generic debugging steps.
Code Reviews: Reference your style guides, architectural decisions, and best practices during reviews without manually looking them up.
Start with the hosted Biel.ai service for immediate setup, or run your own instance if you need complete control over your documentation access.
The self-hosted option includes Docker deployment and can integrate with your existing documentation infrastructure without sending data to external services.
No SDK integrations, no workflow changes, no learning curve. Your existing AI tools just get better at their job.
The difference between generic AI assistance and product-aware AI assistance is the difference between Stack Overflow and your internal wiki. Both have value, but only one knows your specific context.