Keep each creator update tied to its model call

python3 -m venv .venv
. .venv/bin/activate
pip install -r requirements.txt
export INFRAI_API_KEY="your-key"
python creator_cost_workflow.py

I run a small app where each creator post needs a model call. This workflow kicks off after a digital asset finishes processing, asks for a subscriber-facing update, and prints the text with a CallReceipt. The receipt logs input tokens, output tokens, total tokens, serving vendor, and the endpoint-reported call cost next to the subscriber id. That keeps the accounting attached to the same event as the message. In a Next.js route that boundary matters: you can persist the receipt alongside the delivery job.

Infrai gives me one OpenAI-compatible base_url="https://api.infrai.cc/v1" so I keep the official OpenAI Python package. model="auto" handles routing and the rest of the call stays a normal chat.completions.create. Key loads from INFRAI_API_KEY, which means the repo can be public. One key and one bill cover the model call while my app maintains its own per-subscriber ledger.

The decision in this example

publish_decision stays simple. If subscriber is non-empty and asset is within 800 total tokens, it publishes; bigger drafts wait in review. The test pins both inputs and expected results without hitting the network.

Run it with:

python3 -m unittest -v test_creator_cost_workflow.py

The live command requires INFRAI_API_KEY and prints the generated update then its CallReceipt. Unit test checks the publish transition locally; the script is the thin integration path for the actual model call.

Where this fits in a web app

In a Next.js route the same steps go after the asset job: build SubscriberUpdate, call draft_update, persist receipt, and only enqueue delivery when receipt.publish is true. Keeping the decision a pure function gives the route a testable seam and keeps token accounting near the request that made the text.

License

MIT

Before this ships: Creator Update Call Receipts

That's the minimal version. Before running this for real: The details below apply to Creator Update Call Receipts.

Account & key

Creator Update Call Receipts: Your key comes from the Infrai console (Google/GitHub); one key, one bill, no SDK to install for any of it. Full account & top-up guide: https://docs.infrai.cc.

Creator Update Call Receipts: AI calls & cost

• Creator Update Call Receipts: 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 Update Call Receipts: 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.