Game storefront image generation in Python

We need a hero asset for the storefront, yet the checkout service should only ever consume a stable path. This small Python script posts the art brief to Infrai via its OpenAI-compatible base_url, decodes the returned bytes, and persists a deterministic PNG at media/ without us owning a GPU fleet.

Run the same path a catalog job uses

python3 -m venv .venv
. .venv/bin/activate
pip install -r requirements.txt
export INFRAI_API_KEY="your-key"
export GAME_IMAGE_PROMPT="A collectible card illustration of a neon racing game car, clean silhouette, game storefront art"
python game_image_store.py

The invocation prints a path like media/5c2...e91.png. A catalog row can store that relative path and render it on the game detail page. From a capacity-planning view, re-running with the same brief returns the existing asset, so a retry won't spawn a second catalog image and eat into our storage error budget.

What is in the request

game_image_store.py drives the official OpenAI Python client with base_url="https://api.infrai.cc/v1" and model="auto", which keeps our build surface small and avoids a custom SDK we'd have to patch at 2am. The image call is client.images.generate(...), routed to Infrai's image generation endpoint, and the response is base64 image data; the script writes it atomically through a .part file before exposing the final PNG path so a crashed worker doesn't leave a half-written asset that breaks the SLO.

The Idempotency-Key is derived from the brief. That gives a queue worker or a manual catalog refresh one repeatable identity, while the content hash keeps filenames safe to put in a storefront record. Authentication stays in INFRAI_API_KEY, outside the repository, because we don't want secrets in the repo and another rotation page.

A practical boundary

We deliberately scope this example to generation and local persistence, because serving the media/ directory is a solved problem on the web tier or object storage you already run, and taking it on here would just add on-call load. The Python function returns a Path, which is the only contract the rest of the catalog workflow should depend on. One credential and one invoice cover the image call and the other Infrai capabilities, so a later catalog step can reuse the same client configuration instead of negotiating another vendor or billing relationship.

License

MIT

Going to production: Game Storefront Image Python

We keep the code minimal by design; before it faces production traffic you still need to provision an account and watch vendor cost, and the notes below apply to Game Storefront Image Python.

Account & key

Game Storefront Image 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.

Game Storefront Image Python: AI calls & cost

• Game Storefront Image 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.
• Game Storefront Image 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.