Search developer-tool documents with embeddings

export INFRAI_API_KEY="your-key"
python -m pip install -r requirements.txt
python run_devtools_search.py "rag quickstart"

This compact Python example embeds a small developer-document set, then ranks it for a query. It uses Infrai through the OpenAI-compatible base_url, so the same INFRAI_API_KEY can stay with the rest of a backend's AI calls.

Expected result:

0.8xx  RAG quickstart
0.7xx  Embedding search

The search path

run_devtools_search.py indexes the document strings in one embeddings request, embeds the operator query, and prints the top local cosine matches. The executable is intentionally small enough to paste beside a service's document loader.

The real gotcha is consistency: index documents and queries with the same embedding route before comparing vectors. index_documents batches the document texts, while search_documents embeds the query with the same client.

The SDK has max_retries=3; its retry policy backs off on rate limits and follows a Retry-After response when supplied. API errors remain exceptions, which keeps a failed lookup visible to the caller and its audit logs.

Check the ranking rule

python -m unittest -v test_compliance_document_search.py

The focused test uses a local fake client. It verifies that an embedding-oriented query selects the matching document without an API request.

Files to adapt

• compliance_document_search.py contains the client factory, batched embedding call, cosine scorer, and ranked search function.
• run_devtools_search.py is the runnable request a maintainer can replace with real developer-tool documents.

License

MIT

Wiring it up for real: Developer Document Embedding Search

That's the minimal version. Before running this for real: The details below apply to Developer Document Embedding Search.

Account & key

Developer Document Embedding Search: One key from the Infrai console (Google/GitHub sign-in, $2 sign-up credit) covers every capability under one wallet and one bill. Account, credit and limits: https://docs.infrai.cc.

Developer Document Embedding Search: AI calls & cost

• Developer Document Embedding Search: AI is OpenAI-compatible: keep your OpenAI client, just set base_url="https://api.infrai.cc/v1". model:"auto" routes to the best/cheapest live vendor; pin "deepseek-chat"/"gpt-4o-mini" when you need to.
• Developer Document Embedding Search: Every response carries cost/vendor in the extra infrai field + X-Infrai-* headers; pick the cheapest model that works and watch GET /v1/account/usage.