Support & FAQ

Pura LLM is an OpenAI-compatible gateway that routes your requests to LLM providers through a single API. You manage keys, billing, and quotas in the console while keeping your existing OpenAI SDK integration.

Yes. Change base_url to the Pura gateway URL and use your Pura API key. The request and response format matches the OpenAI API for chat completions and model listing.

Pura LLM infrastructure runs in the EU. Each model is tied to a specific region — see the EU Data Residency page for an up-to-date map of the catalog.

No. We apply a zero retention policy: message content is processed in transit and not written to persistent storage. Only billing metadata (token counts, model, timestamp) is retained.

Create an account, open the console, go to API Keys, and generate a new key. Copy it immediately — it is shown only once. Use it as a Bearer token in the Authorization header.

Use https://llm.puradigital.it/v1 as base_url in the OpenAI SDK or as the prefix for REST calls. Chat completions go to /v1/chat/completions.

The Pura LLM model catalog is exposed via GET /public/models (no auth) or GET /v1/models (authenticated). Models, providers, and capabilities are listed in the console Models page and documentation.

Usage is metered by tokens per model. Costs are deducted from your wallet in credits. Exact rates depend on the model and your plan — check the Billing section in the console.

Requests are rejected once your balance is insufficient. Top up via the Billing page or set budget alerts on individual API keys to avoid unexpected interruptions.

Yes. Each key supports rate limits (RPM/TPH), daily cost caps, and optional expiry. Useful for separating production, staging, and client projects.

Yes. Pass stream=true in chat completion requests or use stream=True in the OpenAI Python SDK. Streaming uses the same endpoint and authentication.

Any tool that supports a custom OpenAI base URL works. Point it to the Pura gateway and supply your Pura API key.

Check that the Authorization header is Bearer YOUR_KEY, the key is active, and it has not been revoked. Expired sessions in the console require a new login and a valid key.

You hit a rate limit on your API key or account. Review RPM/TPH settings on the key or wait for the limit window to reset.

The upstream provider or Pura LLM gateway may be temporarily unavailable, or the model name may be misconfigured. Verify the model ID in the catalog and retry.

For models with function_calling capability in the catalog, yes. Pass tools and tool_choice as in the OpenAI API. Capability flags are listed per model in the Models page.

If the model supports vision or pdf_input in the catalog, send multimodal content in the messages array using the OpenAI format (image_url, file references as supported by the provider).

Yes. B2B customers can request a Data Processing Agreement. We act as processor; you remain controller of the data you send through the API.

The Pura LLM console provides Logs, Monitoring, and Dashboard views with token usage, cost, and per-model breakdowns.

Replace OPENAI_BASE_URL with the Pura gateway URL, swap the API key, and verify model IDs match those in the Pura catalog. No changes to message structure are required for standard chat flows.