# OmniRelay Documentation > One OpenAI-compatible API for every model: keys, routing, usage and billing in one gateway. ## V1 Documentation ### V1 Documentation - [OmniRelay Documentation](https://omnirelay.ai/docs/index.md): One OpenAI-compatible API for every model. Create a key, point your SDK at OmniRelay, and track usage and cost in one console. ### Quick Start - [Creating an account](https://omnirelay.ai/docs/quickstart/joining-the-beta.md): Open an OmniRelay account. While the closed beta gate is on, you need an invite code or an approved waitlist request. - [First API call](https://omnirelay.ai/docs/quickstart/first-api-call.md): Discover your callable models and make a live text call using an OmniRelay key. - [Set up with an AI agent](https://omnirelay.ai/docs/quickstart/ai-agent-setup.md): Copy one prompt into Claude Code, Codex, Cursor or another coding agent and let it add OmniRelay to your project. ### User manual - [User manual](https://omnirelay.ai/docs/user-manual.md): Accounts, keys, model selection, playground, API calls, usage, billing and troubleshooting. ### Overview - [Unified gateway](https://omnirelay.ai/docs/overview/unified-gateway.md): What OmniRelay does between your request and the model. - [Authentication](https://omnirelay.ai/docs/overview/authentication.md): Use Bearer authentication with inference API keys or separate management keys. - [Rate limits](https://omnirelay.ai/docs/overview/rate-limits.md): Request rate, concurrency, key IP allowlists and payload limits enforced by the gateway. ### LLMs - [List models](https://omnirelay.ai/docs/llms/listing-models.md): Discover the models enabled in your workspace. - [Chat completions](https://omnirelay.ai/docs/llms/chat-completions.md): Call enabled text models with JSON or OpenAI-compatible streaming responses. - [Responses and embeddings](https://omnirelay.ai/docs/llms/native-protocols.md): Stateless Responses, typed stream events and reviewed embedding models. - [Request metadata](https://omnirelay.ai/docs/llms/request-metadata.md): Opt in to a per-request report of what the gateway decided: the model and provider that served you, the mode, the attempts and the recorded tags. ### Images - [Image generation and editing](https://omnirelay.ai/docs/images/image-generation.md): Create and edit images with OpenAI-compatible endpoints, billed per image by size tier. ### Videos - [Video generation](https://omnirelay.ai/docs/videos/video-generation.md): Generate videos from a text prompt with an asynchronous task API, billed by the output tokens generated. - [Video understanding](https://omnirelay.ai/docs/videos/video-understanding.md): Ask a Gemini model about a video's picture and sound through the Gemini-native generateContent API, billed by tokens. ### General - [Usage and billing](https://omnirelay.ai/docs/general/usage-billing.md): How requests are metered, priced and recorded. - [Model access and request diagnostics](https://omnirelay.ai/docs/general/provider-integration.md): Confirm live model access, inspect request records and understand holds and settlement. - [Tags](https://omnirelay.ai/docs/general/tags.md): Label calls with key/value tags such as client=acme or env=prod to split usage and cost per customer, project or environment. - [Management API](https://omnirelay.ai/docs/general/management-api.md): Provision API keys and pull usage from scripts, with a management key that cannot call models. - [Data handling](https://omnirelay.ai/docs/general/data-handling.md): What OmniRelay records about your requests and account, what it never stores, where it lives and how long it is kept. - [Errors](https://omnirelay.ai/docs/general/errors.md): Error format and every error code returned by the /v1 API. - [Service boundaries](https://omnirelay.ai/docs/general/service-boundaries.md): Capabilities, activation and limits of the current live model service. - [OmniRelay console assistant](https://omnirelay.ai/docs/general/console-assistant.md): Use the signed-in console assistant for product guidance and understand its scope and data handling. ## V1 API Reference ### Service - [Health check](https://omnirelay.ai/docs/api-reference/service/health-check.md): Returns the public service status. A healthy response only means the application is answering; it does not verify a model provider, payments, or backups. ### Models - [List models](https://omnirelay.ai/docs/api-reference/models/list-models.md): Returns callable real models filtered by workspace, key allowlist and available runtime/channels. Use your own response for IDs, capabilities, limits, reasoning efforts and prices. Catalog visibility does not grant access. ### Chat - [Create chat completion](https://omnirelay.ai/docs/api-reference/chat/create-chat-completion.md): Calls an enabled model with its protocol-specific reviewed capabilities. Query GET /v1/models for tools, JSON formats, sampling, reasoning and combination restrictions. Capability review covers every callable route. New combinations are documentation-based, pending individual live validation. Reserv… ### Management - [List API keys](https://omnirelay.ai/docs/api-reference/management/list-api-keys.md): Returns only the owner's API keys, including revoked entries. No full credentials are returned. - [Create API key](https://omnirelay.ai/docs/api-reference/management/create-api-key.md): Creates an API key for the owner after verified workspace activation. Otherwise returns 403 activation_required without creating a key. The key is returned once. Maximum 20 active API keys per owner. Amounts are integer micro-USD. - [List key budgets](https://omnirelay.ai/docs/api-reference/management/list-key-budgets.md): Returns owner key budgets and effective remaining allowance. Amounts are micro-USD; reset timestamps are Unix milliseconds, with UTC calendar resets. - [Revoke API key](https://omnirelay.ai/docs/api-reference/management/revoke-api-key.md): Revokes an owner key; its next authentication fails. This does not delete historical usage. - [Update API key](https://omnirelay.ai/docs/api-reference/management/update-api-key.md): Updates supplied fields on an owner key. Omitted fields are preserved. Use the budget endpoint for spending controls. - [Rotate API key](https://omnirelay.ai/docs/api-reference/management/rotate-api-key.md): Rotates an active owner key; preserves ID, restrictions, expiry and spending counters. The old credential stops working; the new key is returned once. Limited to 10 rotations per owner per hour. - [Update key budget](https://omnirelay.ai/docs/api-reference/management/update-key-budget.md): Updates spending controls. Changing budget does not refill existing remaining allowance. To refill, explicitly set remaining. Clearing budget requires resetPeriod=none and remaining=null. Periodic budgets must be positive. - [Read usage summary](https://omnirelay.ai/docs/api-reference/management/read-usage-summary.md): Returns UTC usage totals and breakdowns. scope is workspace for owners with members.read permission, otherwise user. Also subject to a shared monitoring limit of 240 requests per owner per minute. - [List request records](https://omnirelay.ai/docs/api-reference/management/list-request-records.md): Returns newest-first request records, including failures. Pass next as before to fetch another page; next=null ends pagination. Uses the same permission scope and monitoring limit as usage summary. ### Responses - [Create responses](https://omnirelay.ai/docs/api-reference/responses/create-response.md): Calls an enabled model with its protocol-specific reviewed capabilities. Query GET /v1/models for tools, JSON formats, sampling, reasoning and combination restrictions. Capability review covers every callable route. New combinations are documentation-based, pending individual live validation. Reserv… ### Embeddings - [Create embeddings](https://omnirelay.ai/docs/api-reference/embeddings/create-embeddings.md): Input-only embeddings for models with capabilities.embeddings=true. No embedding model is currently enabled. A reviewed model, dimensions, aggregate input bound and input price are required before activation. Float vectors only; batch size and adjustable dimensions depend on the model. ### Images - [Create images](https://omnirelay.ai/docs/api-reference/images/create-images.md): Generates images from a text prompt with an image model (capabilities.image_generation=true). Billed per image returned at the tier of the requested `size`: longest side up to 1024 px is 1K, up to 2048 px is 2K, up to 4096 px is 4K; an omitted or `auto` size is billed as 2K. Quality does not change… - [Edit images](https://omnirelay.ai/docs/api-reference/images/edit-images.md): Edits or extends one or more uploaded images (multipart/form-data, fields `image` or `image[]`, optional `mask`). Billing, size tiers, `n`, `b64_json` output and the long-call behaviour are the same as Create images. The whole request may be up to 20 MiB. ### Videos - [Create video task](https://omnirelay.ai/docs/api-reference/videos/create-video-task.md): Submits a text-to-video task and returns its ID at once; poll the task until it finishes. Billed by output tokens generated (USD 25 per million list price): width x height x (24 x seconds + 1) / 1024. The cost of the largest frame at the requested resolution and duration, plus 10 percent, is held at… - [Get video task](https://omnirelay.ai/docs/api-reference/videos/get-video-task.md): Returns a task created with the same user. Reading a finished task settles its charge if that has not happened yet; tasks are settled when they finish whether or not they are read. ## OpenAPI Specs - [openapi](/docs/openapi.json) > The links below point to documentation indexes. Follow each `/_llms/` index recursively until you reach documentation pages. ## Indexes - [Chinese (41 pages)](https://omnirelay.ai/docs/_llms/cn.txt): V1 文档, V1 API 参考. - [Spanish (3 pages)](https://omnirelay.ai/docs/_llms/es.txt): Documentación V1. - [French (3 pages)](https://omnirelay.ai/docs/_llms/fr.txt): Documentation V1. - [German (3 pages)](https://omnirelay.ai/docs/_llms/de.txt): V1-Dokumentation. - [Brazilian Portuguese (3 pages)](https://omnirelay.ai/docs/_llms/pt-br.txt): Documentação V1. - [Japanese (3 pages)](https://omnirelay.ai/docs/_llms/jp.txt): V1 ドキュメント. - [Korean (3 pages)](https://omnirelay.ai/docs/_llms/ko.txt): V1 문서. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.