Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.clearmaas.com/llms.txt

Use this file to discover all available pages before exploring further.

What’s available

ClearMaas routes to nine upstream providers. Models are addressed by provider-prefixed ID:
PrefixProviderExamples (call /v1/models for the live catalog)
openai/OpenAIopenai/gpt-4o-mini, openai/gpt-5, openai/o3-mini, openai/gpt-image-1, openai/tts-1
anthropic/Anthropicanthropic/claude-sonnet-4.6, anthropic/claude-opus-4.7
google/Google Geminigoogle/gemini-2.5-flash, google/gemini-3-pro-preview, google/imagen-4.0-generate-001
deepseek/DeepSeekdeepseek/deepseek-chat, deepseek/deepseek-reasoner
grok/xAI Grokgrok/grok-4-fast-reasoning, grok/grok-imagine-image
qwen/Alibaba Qwenqwen/qwen3.6-plus, qwen/qwen3-vl-235b-a22b-thinking, qwen/qwen3-max
kimi/Moonshot Kimikimi/kimi-k2.5, kimi/kimi-k2.6
minimax/MiniMaxminimax/minimax-m2.7, minimax/minimax-m2.5-highspeed
kling/Kling (video)kling/kling-v3-omni, kling/kling-v2-6 — see Kling Video
Whichever model you’d call against the provider’s own API, you can call against ClearMaas under the prefixed ID. The catalog is exactly what each upstream publishes.

List all models programmatically

curl https://api.clearmaas.com/v1/models \
-H "Authorization: Bearer sk-clearmaas-..."

Response shape

{
  "object": "list",
  "data": [
    {
      "id": "openai/gpt-4o",
      "object": "model",
      "created": 1626777600,
      "owned_by": "openai",
      "supported_endpoint_types": ["openai"]
    },
    {
      "id": "anthropic/claude-sonnet-4.6",
      "object": "model",
      "owned_by": "anthropic",
      "supported_endpoint_types": ["anthropic", "openai"]
    }
  ]
}
The list only includes models your account has access to and that have a pricing ratio configured. See the API Reference / Models for the full endpoint spec.

Get one model

curl https://api.clearmaas.com/v1/models/openai/gpt-4o-mini \
-H "Authorization: Bearer sk-clearmaas-..."

Naming conventions

Model IDs in ClearMaas are provider-prefixed by default. The prefix tells you (and your client-side logs) exactly which provider will serve the call before the request even leaves your machine. The bare upstream name (e.g. gpt-4o-mini) may also work if your administrator has configured it as an alias on a channel, but the default listing in /v1/models always uses the prefixed form — code written against the prefixed names is portable across deployments.

Picking by capability

You need…Try
Cheapest chatopenai/gpt-4o-mini, google/gemini-2.5-flash-lite
Strongest reasoningopenai/o3-mini-high, openai/gpt-5-pro, anthropic/claude-opus-4.7
Long context + cheapgoogle/gemini-2.5-flash
Web search built-inopenai/gpt-4o-search-preview (chat completions), or any Grok model on /v1/responses with the web_search tool — see Web search
Vision inputopenai/gpt-4o-mini, google/gemini-2.5-flash, grok/grok-4-fast-reasoning
Image generationopenai/gpt-image-1-mini, google/imagen-4.0-fast-generate-001, grok/grok-imagine-image
Auto-pick cheapestclearmaas/auto — see Auto Router
Video generationkling/kling-v3-omni, kling/kling-v3 — see Kling Video