# AlphaNeural API

> One API key and one prepaid balance for every model in the AlphaNeural catalogue, behind an OpenAI-compatible chat endpoint that also keeps conversations server-side (threads), routes between models under a spend ceiling, and reports per-task cost traces.

Primary base URL: https://backend.alfnrl.io/v1 (chat completions with threads, routing and traces). Raw, stateless OpenAI-compatible gateway: https://proxy.alfnrl.io/v1. Both take the same API key, sent as "Authorization: Bearer <key>".

## Start

- [AlphaNeural API](https://app.alphaneural.io/docs/index.md): One API key and one balance for every model in the catalogue, behind an OpenAI-compatible endpoint that also keeps your conversations, routes between models under a spend ceiling, and itemises what each task cost.
- [Quickstart](https://app.alphaneural.io/docs/quickstart.md): Get a key, make a first call, and continue the same conversation on a different model, in three steps.
- [Authentication](https://app.alphaneural.io/docs/authentication.md): Where API keys come from, which kinds work, and how to send one.
- [Billing](https://app.alphaneural.io/docs/billing.md): Per-token prices with the margin already included, one prepaid balance, what a 402 means, and where to check what you spent.

## API

- [Chat completions](https://app.alphaneural.io/docs/chat-completions.md): POST /v1/chat/completions: the OpenAI request and response, plus the fields that add threads, routing and cost traces.
- [Threads and memory](https://app.alphaneural.io/docs/threads.md): Keep a conversation on the server, continue it on any model, and read it back.
- [Routing and spend ceilings](https://app.alphaneural.io/docs/routing.md): Send alphaneural/auto, or your own ladder of models, and the server tries a cheap model first, checks its answer, escalates only when the check fails, and never spends past your ceiling.
- [Cost traces](https://app.alphaneural.io/docs/traces.md): Tag requests with a task id, list your tasks, and read back what each step cost, with an estimate of what one baseline model would have cost instead.
- [Models](https://app.alphaneural.io/docs/models.md): List the models your key can call, with the price you are billed, context windows and input types, and the catalogue query behind the pricing page.
- [Streaming](https://app.alphaneural.io/docs/streaming.md): Server-sent events from the chat endpoint, and what is saved to the thread when a stream is cut short.
- [Errors](https://app.alphaneural.io/docs/errors.md): Every status and message the API returns, and which errors are passed through from the model provider unchanged.
- [The raw gateway](https://app.alphaneural.io/docs/gateway.md): proxy.alfnrl.io, the stateless OpenAI-compatible gateway, and when to use it instead of the primary endpoint.

## Guides

- [Use the OpenAI SDK](https://app.alphaneural.io/docs/guides/openai-sdk.md): Point the official Python or TypeScript SDK at AlphaNeural, and reach threads, routing and traces from it.

## Optional

- [All pages in one file](https://app.alphaneural.io/docs/llms-full.txt): every page above, concatenated.