# Celeris

> Celeris is an OpenAI-compatible inference API for celeris-1, a diffusion
> LLM tuned for fast, short, structured responses (classification, extraction,
> judging, query rewriting) with latency in the tens of milliseconds.

- Console: https://console.celeris.ai
- Docs: https://docs.celeris.ai

## Docs

- [Command-line interface](https://docs.celeris.ai/cli.md): Install the celeris CLI and call the API from your shell, with streaming, piping, and JSON output.
- [Companies and workspaces](https://docs.celeris.ai/organizations.md): How company context, workspace access, roles, and the flat workspace switcher fit together.
- [Quickstart](https://docs.celeris.ai/quickstart.md): Create an API key and make your first Celeris request in about five minutes.
- [Account & sign-in](https://docs.celeris.ai/account.md): Signing up, verification codes, password resets, workspace invites, and what to do when an email doesn't arrive.
- [Console & billing](https://docs.celeris.ai/console.md): The Customer Console, covering API keys, the Playground, usage, credits, Commercial access, and your team.
- [Activation waitlist](https://docs.celeris.ai/waitlist.md): What the workspace activation waitlist means and how to continue after your email arrives.
- [Authentication](https://docs.celeris.ai/authentication.md): How Celeris API keys work, from creating and sending them to rotating and revoking them.
- [Models](https://docs.celeris.ai/models.md): The models Celeris serves, what each one is for, and how model routing works.
- [Making requests](https://docs.celeris.ai/making-requests.md): OpenAI compatibility, supported parameters, browser calls, and response behavior.
- [Image input](https://docs.celeris.ai/images.md): Send images alongside text to celeris-1 using OpenAI-compatible content parts.
- [Errors](https://docs.celeris.ai/errors.md): Every error the API returns, with status, body, causes, and how to handle each one.
- [Rate limits](https://docs.celeris.ai/rate-limits.md): What's limited, 429 handling, and how to back off well.
- [Prompt engineering](https://docs.celeris.ai/prompting.md): How to prompt a diffusion LLM for fast, structured, reliable outputs.
- [Agentic workloads](https://docs.celeris.ai/agentic-workloads.md): Request patterns celeris-1 excels at, such as classify, extract, judge, and rewrite inside agent loops.
- [Measuring latency](https://docs.celeris.ai/measuring-latency.md): Understand end-to-end, first-token, and response-head latency.
- [Going to production](https://docs.celeris.ai/going-to-production.md): A production-readiness checklist for keys, retries, timeouts, concurrency, observability, and billing.
- [Pricing](https://docs.celeris.ai/pricing.md): Token rates, what is and isn't charged, estimating costs, and how prepaid credits work.
- [Availability](https://docs.celeris.ai/availability.md): Where Celeris is available today, and how to hear from us when it reaches you.
- [Security and data handling](https://docs.celeris.ai/security.md): How Celeris protects API credentials, separates workspaces, and records usage.
- [Known issues](https://docs.celeris.ai/known-issues.md): Current model-output issues, examples, and safe mitigations.
- [Glossary](https://docs.celeris.ai/glossary.md): The terms used across Celeris, covering the models, the API, the Console, and billing.
- [Troubleshooting & support](https://docs.celeris.ai/support.md): Diagnose common failures fast, and what to include when you contact support.
- [API reference](https://docs.celeris.ai/api-reference.md): The Celeris HTTP API, with endpoints, request fields, response schema, and headers.