Examples

Short, runnable examples for each feature. The curl examples assume ACS_API_KEY + ACS_API_BASE are exported (see Quick start); the Python examples use the same openai SDK client.

Each page below shows one shared example response after the curl and Python snippets — both calls return the same JSON; the SDK wraps it in typed objects. Always-null fields (service_tier, system_fingerprint, kv_transfer_params, etc.) are dropped from the shown responses for brevity.

Your own output won't match token-for-token: the displayed requests don't pin a seed, so under the default temperature=1.0 sampling varies. Add a seed to reproduce a specific run, or temperature=0 for greedy decoding.

Guided walkthrough

The Colab workshop introduces completions, sampling, logprobs, reading activations, and building a steering vector in your browser. Exercises and an instructor solutions notebook are in our workshop repo.

Token-level inspection

• logprobs — top-k logprobs per generated token.
• prompt_logprobs — top-k logprobs at each prompt position (gotchas around rank-vs-actual-token).
• echo — include the prompt in the response.

Streaming + concurrency

• stream — SSE token-by-token.
• Batch rollouts — 16 concurrent per key, using asyncio.gather.

Recovery patterns

• Cold-boot waiting — keep one request open through a large-model startup.
• Budget-cap recovery — handle 429 budget_exceeded (the non-retryable kind).

Evaluation

• Evaluate with Inspect — run base-model evals through Inspect's openai-api-completions provider (raw prompts, no chat template).

• Completion jobs — submit once, reconnect, and retrieve the actual result status.