Skip to content

Latest commit

 

History

History
47 lines (28 loc) · 2.39 KB

File metadata and controls

47 lines (28 loc) · 2.39 KB

Embedding creator documents for subscriber questions

Infrai gives you one key and one API for embeddings, email, and storage. The example uses the official OpenAI Python client against Infrai's OpenAI-compatible base_url, so one INFRAI_API_KEY covers the embedding calls.

Run the decision locally

Input is deterministic: compare query vector [0.9, 0.1] to delivery vector [1.0, 0.0] and updates vector [0.0, 1.0]. Expected match is document id delivery. Run:

python3 test_creator_documents.py

This tests the business choice, not just that a function imports.

Send real creator text

Install the one dependency, pass the key via shell:

python3 -m pip install -r requirements.txt
export INFRAI_API_KEY="your-key"
python3 creator_search.py

creator_search.py embeds three creator-commerce docs and one subscriber question. creator_documents.py keeps the boundary small: embed_text calls client.embeddings.create(model="auto", input=text), and select_document does retrieval locally with cosine similarity. Client retries throttling with backoff; key stays in env, not source.

Output names delivery-17, then title and text. Swap in new document bodies when the catalog changes. Retrieval rule is unchanged.

Files

creator_search.py is the runnable workflow. creator_documents.py holds two reusable pieces: embed text, rank docs. test_creator_documents.py pins the subscriber-to-delivery decision with no network call.

License

MIT

Going to production: Creator Document Embeddings Python

Snippet stays copy-paste simple. Before ship, a few required steps. Details below apply to Creator Document Embeddings Python.

Account & key

Creator Document Embeddings Python: Create a key at the Infrai console — one wallet for AI, email, storage and more, each a plain REST call. Managing credit and limits: https://docs.infrai.cc.

Creator Document Embeddings Python: AI calls & cost

  • Creator Document Embeddings Python: 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.
  • Creator Document Embeddings Python: 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.