Endpointer
An AI-assisted search tool that indexes vendor API documentation and gives engineers and coding agents direct access to integration context.
About the project
I designed and built Endpointer to turn scattered vendor documentation and API specifications into searchable engineering context. Its ingestion pipeline creates structured operation and field indexes, preserves source references, and exposes search through a Python CLI and read-only Model Context Protocol tools. Engineers and AI assistants can retrieve relevant API details without repeating manual searches across documentation sites.
Interface views
Select an image to view it at full size.
1. The question: use Endpointer through Codex to find an API operation in locally indexed documentation.
Image provided
2. The retrieval: search_docs and search_endpoints locate relevant documentation; get_operation retrieves the selected contract.
Image provided
3. The answer: integration guidance based on the retrieved documentation, with parameters, response fields, documented errors, and remaining questions.
Image provided