# OrcaRouter

## Docs

- [Audio input](https://docs.orcarouter.ai/advanced/audio-input.md): Send audio to a multimodal Gemini model via /v1/chat/completions or the native /v1beta/ path.
- [Reasoning](https://docs.orcarouter.ai/advanced/reasoning.md): Reasoning models across all providers, with one unified effort syntax.
- [Streaming](https://docs.orcarouter.ai/advanced/streaming.md): Server-Sent Events for incremental output — Chat Completions, Responses, and Anthropic Messages.
- [Structured outputs](https://docs.orcarouter.ai/advanced/structured-outputs.md): Constrain model output to JSON or a JSON schema via response_format.
- [Tool calling](https://docs.orcarouter.ai/advanced/tool-calling.md): Pass tools / function declarations and let the model call them.
- [Vision (image input)](https://docs.orcarouter.ai/advanced/vision.md): Send images to vision-capable models via image_url content parts.
- [Web search](https://docs.orcarouter.ai/advanced/web-search.md): Let the model search the web before answering — via search-preview models or web_search_options.
- [Generate speech](https://docs.orcarouter.ai/api-reference/audio/generate-speech.md)
- [Create a chat completion](https://docs.orcarouter.ai/api-reference/chat/create-a-chat-completion.md): OpenAI-compatible chat completions. Supports:
- [Gemini native generate](https://docs.orcarouter.ai/api-reference/gemini-native/gemini-native-generate.md): Direct passthrough to Google's Gemini API. The `model_path` parameter is `<model>:<action>`, for example `google/gemini-2.5-flash:generateContent` or `google/gemini-2.5-flash:streamGenerateContent`.
- [Create an image](https://docs.orcarouter.ai/api-reference/images/create-an-image.md): Supported across providers: OpenAI (`gpt-image-1`, `gpt-image-1-mini`, `gpt-image-1.5`), Google Imagen (`imagen-4.0-{fast,standard,ultra}-generate-001`), and xAI Grok (`grok-imagine-image`, `grok-imagine-image-pro`).
- [Create a message](https://docs.orcarouter.ai/api-reference/messages/create-a-message.md)
- [Get model details](https://docs.orcarouter.ai/api-reference/models/get-model-details.md)
- [List available models](https://docs.orcarouter.ai/api-reference/models/list-available-models.md)
- [Create a response](https://docs.orcarouter.ai/api-reference/responses/create-a-response.md): OpenAI Responses API. Supported on **OpenAI models** and **xAI Grok models** — xAI's Agent Tools API uses the same shape, so any Grok model accepts `tools: [{"type": "web_search"}]` / `x_search` / `code_interpreter` here. Anthropic and Gemini targets return `not_implemented` because their native sha…
- [Fetch video task status](https://docs.orcarouter.ai/api-reference/video/fetch-video-task-status.md): When `status` is `SUCCESS`, `data.result_url` carries the MP4 (upstream URL, short TTL — download promptly).
- [Submit a video task](https://docs.orcarouter.ai/api-reference/video/submit-a-video-task.md): Submit an async video-generation task.
- [Anthropic SDK](https://docs.orcarouter.ai/compatibility/anthropic-sdk.md): Point the official Anthropic SDK at OrcaRouter to call Claude models via /v1/messages.
- [Frameworks & Integrations](https://docs.orcarouter.ai/compatibility/frameworks.md): LangChain, Vercel AI SDK, Claude Code, Codex CLI, and any tool that lets you override the OpenAI or Anthropic base URL.
- [Google GenAI SDK](https://docs.orcarouter.ai/compatibility/google-genai-sdk.md): Point the google-genai SDK at OrcaRouter to call Gemini models via /v1beta/.
- [OpenAI SDK](https://docs.orcarouter.ai/compatibility/openai-sdk.md): Change one line to use OrcaRouter with the official OpenAI SDK.
- [Guardrails](https://docs.orcarouter.ai/features/guardrails.md): Workspace-scoped content-policy engine. Attach once, every bound key is screened on the next call — block, mask, or flag prompts and responses with no redeploy and no SDK change.
- [Prompts](https://docs.orcarouter.ai/features/prompts.md): Workspace-scoped prompt registry. Edit once, every bound key picks up the new version on the next call — no redeploy, no SDK change.
- [Get an API key](https://docs.orcarouter.ai/getting-started/get-api-key.md): Sign up, create an OrcaRouter API key, and authenticate.
- [Models](https://docs.orcarouter.ai/getting-started/models.md): Discover and pick from the catalog. Models are provider-prefixed (openai/, anthropic/, google/, deepseek/, grok/, qwen/, kimi/, minimax/, z-ai/, kling/, byteplus/).
- [Quickstart](https://docs.orcarouter.ai/getting-started/quickstart.md): Make your first OrcaRouter call in under a minute.
- [Welcome to OrcaRouter](https://docs.orcarouter.ai/introduction.md): One API. Multi-provider. Zero markup.
- [Kling Video](https://docs.orcarouter.ai/kling-video/overview.md): Async video generation via /v1/video/generations — text-to-video, image-to-video, and Omni-Video (multi-source / multi-shot) on Kling models.
- [Anthropic native HTTP](https://docs.orcarouter.ai/native-formats/anthropic.md): Raw HTTP shape for POST /v1/messages — content blocks, tool use, named SSE events.
- [Gemini native HTTP](https://docs.orcarouter.ai/native-formats/gemini.md): Direct passthrough to /v1beta/models/{model}:generateContent — full Gemini protocol fidelity.
- [OpenAI-compatible HTTP](https://docs.orcarouter.ai/native-formats/openai-compat.md): Raw HTTP shape for /v1/chat/completions and /v1/responses without an SDK.
- [Billing & usage](https://docs.orcarouter.ai/operations/billing-and-usage.md): How OrcaRouter computes charges, the usage object, and OrcaRouter-specific response headers.
- [Data Handling](https://docs.orcarouter.ai/operations/data-handling.md): What OrcaRouter stores, what we don't, and how long we keep it.
- [Errors](https://docs.orcarouter.ai/operations/errors.md): HTTP status codes, error envelope, and how to handle upstream vs OrcaRouter errors.
- [Rate Limits](https://docs.orcarouter.ai/operations/rate-limits.md): OrcaRouter rate-limits per workspace. How to handle 429s.
- [Zero Data Retention](https://docs.orcarouter.ai/operations/zero-data-retention.md): Prompts and outputs are processed in memory and never written to disk.
- [Images](https://docs.orcarouter.ai/other-apis/images.md): Generate images via /v1/images/generations or via chat completions with imagine-capable models.
- [Text-to-speech](https://docs.orcarouter.ai/other-apis/tts.md): Generate speech audio from text via /v1/audio/speech (OpenAI shape) or Gemini's native /v1beta/.
- [Auto Router](https://docs.orcarouter.ai/routing/auto-router.md): orcarouter/auto — let OrcaRouter pick the cheapest live model per request.
- [Model Fallbacks](https://docs.orcarouter.ai/routing/model-fallbacks.md): Chain multiple models — if one fails, try the next.
- [Named Routers](https://docs.orcarouter.ai/routing/named-routers.md): Save a routing strategy and invoke it by name.
- [Response Headers](https://docs.orcarouter.ai/routing/response-headers.md): X-Orca-* headers that tell you how your request was routed.
- [Seedance Video](https://docs.orcarouter.ai/seedance-video/overview.md): Async video generation via /v1/video/generations — text-to-video, image-to-video, multimodal reference (image/video/audio), and video editing on Seedance models.

## OpenAPI Specs

- [openapi](https://docs.orcarouter.ai/openapi.yaml)