Crazyrouter API Document
🇺🇸 English
  • 中文
  • 🇺🇸 English
🇺🇸 English
  • 中文
  • 🇺🇸 English
🇺🇸 English
  • 中文
  • 🇺🇸 English
API 参考
快速开始接入教程官网模型与价格控制台
API 参考
快速开始接入教程官网模型与价格控制台
  1. Decisions API (TypeSafe JEV)
  • 默认模块
    • Crazyrouter Overview
      • Introduction
      • API request
      • API Quick Start Guide
      • API Endpoint
      • Contact us
    • Chat
      • ChatGPT
        • ChatGPT Audio
          • whisper-1
          • Create Translation (whisper-1, audio to English, untested)
        • ChatGPT Chat
          • Chat Completion Object
          • Chat Completion Chunk Object
          • Create Chat Completion (Streaming)
          • Create Chat Completion (Non-Streaming)
          • Create Chat Vision (Streaming)
          • Create Chat Vision (Streaming) base64
          • Create Chat Vision (Non-Streaming)
          • Create Chat Image Generation (Non-Streaming)
          • Function calling
          • Official N Test
          • Create Chat Function Call
          • Create Structured Output
          • Control Reasoning Model Effort Level
          • Create Chat Completion qwen-mt-turbo
          • Create Chat Completion DeepSeek V3.1 Thinking Level (Streaming)
          • DeepSeek OCR Recognition
        • ChatGPT Completions
          • Completion Object
        • ChatGPT Embeddings
          • Embedding Object
          • Create Embedding
        • Web Search
          • web search
        • ChatGPT Moderations
          • Content Moderation (Moderations)
      • Anthropic Claude
        • Chat Completion Object
        • Chat Completion Chunk Object
        • Count Input Tokens count_tokens [Native Format]
        • Create Chat Completion (Streaming) [Native Format]
        • Create Function Call (Streaming) [Native Format]
        • Create Thinking Chat [Native Format]
        • Create Thinking Chat
        • Create Chat Completion (Streaming)
        • Create Chat Completion (Non-Streaming)
        • Create Chat Vision (Streaming)
        • Create Chat Vision (Non-Streaming)
        • PDF Support [Native Format]
      • Google Gemini
        • Gemini Native Format
          • Create Context Cache cachedContents [Native Format]
          • Text Generation
          • Text Generation Streaming
          • Text Generation Thingking+streaming
          • Image Generation gemini-2.5-flash-image
          • Image Generation gemini-2.5-flash-image Control Aspect Ratio +imageSize
          • Image Understanding
          • Formatted Output
          • Function Call
          • Document Understanding
          • URL context
          • Code Execution
          • Video Understanding
          • URL context
          • Video Understanding-url Native Format
          • Audio Understanding
          • Text Generation gemini-3-pro-preview:generateContent
        • Chat-compatible Format
          • Gemini Image Generation API [Chat Compatible Format]
          • Chat API [Chat Compatible Format]
          • Chat API - Thinking 1 [Chat Compatible Format]
          • Chat API - Thinking 2 [Chat Compatible Format]
          • Vision API [Chat Compatible Format]
          • Chat + File Reading API [Chat Compatible Format]
    • Chat (Responses)
      • Responses API vs Chat API
      • Create Function Calling
      • Create Model Response (Streaming)
      • Create Model Response (Control Thinking Length)
      • web search
      • Create Model Response (GPT-5 Enable Thinking)
    • Image Generation Models
      • Image Object
      • Midjourney
        • Submit Video Task (Image-to-Video)
        • Submit Simple Change Task (String Form)
        • Get Task Image (Gateway-Hosted Copy)
        • Submit Edits Task (Inpainting)
        • Submit Shorten Task (Prompt Shortening)
        • Submit Change Task (U/V/R Actions)
        • InsightFace Face Swap
        • upload image
        • upload Imagine task
        • Query Task Status by Task ID
        • Query Tasks by ID List
        • Get Task Image Seed
        • Execute Action
        • Submit Blend Task
        • Sumit Describe task
        • Sumit Modal
      • GPT Image (gpt-image-2)
        • Create gpt-image-1
        • Create Image (gpt-image-2, OpenAI images/generations format)
        • Edit Image (gpt-image-2, mask / multi-image reference)
      • Jimeng Image
        • Create Image
        • Edit Image
      • Doubao Seedream
        • doubao-seedream-3-0-t2i-250415
        • doubao-seededit-3-0-i2i-250628
        • doubao-seedream-4-0-250828 Text-to-Image
        • doubao-seedream-4-0-250828 Image-to-Image
        • doubao-seedream-4-0-250828 Multi-Image to Image
        • doubao-seedream-4-5-251128 Text-to-Image (text-only input, single image output)
        • doubao-seedream-4-5-251128 Text+Image to Image (single image in, single image out)
        • doubao-seedream-4-5-251128 Multi-Image Fusion (multiple inputs, single output)
        • doubao-seedream-4-5-251128 Image Set Output (multi-image output)
        • doubao-seedream-4-5-251128 Single Image to Image Set
        • doubao-seedream-4-5-251128 Multi-Reference to Image Set
      • Qwen-Image
        • qwen-image-edit-2509
    • Video Models
      • Veo
        • Unified Video Format
          • Unified Video API Task Status
          • Create Video with Image
          • Create Video (Reference Image)
        • OpenAI Video Format
      • Kling
        • Callback Protocol
        • Omni Video omni-video (kling-v3-omni / kling-video-o1)
        • Image Generation
        • Text-to-Video
        • Image to Video
        • Query Task (free)
        • Multi-Image Reference to Video
        • Lip Sync
        • Extend Video
        • Video Effects
      • Jimeng Video
        • Jimeng Task Status
        • Submit Video Generation Task
        • Query Video Task (free)
      • Hailuo (MiniMax)
        • Hailuo Task Status
        • Submit Video Generation Task
        • Query Video Task Status
        • Image-to-Video
      • Doubao Seedance
        • Text-to-Video Example
        • Image to Video - First Frame
        • seedance-lite First/Last Frame
        • Image-to-Video (base64 encoded)
        • Query Single Task
      • Sora
        • OpenAI Video Format
          • OpenAI Query Task
          • OpenAI Download Video
          • OpenAI Create Video (with image)
          • Create Video from Storyboard
          • OpenAI Create Video with Image (Private Mode)
          • OpenAI Create Video (with Character)
      • Tongyi Wanxiang
        • Generate Video
        • Query Video
      • Unified Video API (all video models)
        • Query Video Task (Unified Format)
        • Submit Video Generation Task (Unified Format)
    • System API
      • List Tokens
      • Create Token
      • Get Models Supported by Token
      • Get Account Info
      • Update Token
      • Delete Token
      • Search Tokens
      • Query Async Task (Image / Video)
      • Get Single Model Info
    • Suno Music
      • Suno Parameters
      • Suno Overview
      • Submit Tasks
        • Generate Song (Inspiration Mode)
        • Generate Song (Custom Mode)
        • Generate Song (Extend Mode)
        • Generate Song (Artist Style)
        • Generate Song (Upload Song for Re-creation)
        • Generate Lyrics
        • Concatenate Songs
      • Query Tasks
        • Batch Get Tasks
        • Query Single Task
        • Get WAV
    • Python
      • Python Speech-to-Text (whisper-1)
      • Python Embeddings
      • Python Function Calling Demo
      • Python LangChain Demo
      • Python llama_index Configuration
      • Python Basic Chat
      • Python Local Image Recognition (Multimodal)
      • Python Remote Image Recognition (Multimodal)
      • Python Image Recognition with Claude
      • Python Library Streaming Output
      • Python requests streaming output demo
      • Python image generation and editing (gpt-image-2)
      • Python official OpenAI library (AutoGPT / LangChain, etc.)
      • Python multi-turn conversation
    • Rerank
      • Rerank
    • PHP
      • PHP Image Editing Demo
    • Node.js
      • Node.js Audio Transcription Demo
    • Plugins & Client Setup Guides
      • Codex
      • N8N
      • Gemini CLI
      • Claude Code Installation and Usage Guide
      • CherryStudio Configuration Guide
      • dify Configuration Guide
      • cline
      • aider
      • Cursor
      • lobechat
      • ChatBox
      • gpt_academic
      • nextchat
      • zotero gpt
      • CLAUDE DEV
      • ChatGPT Sidebar
      • chatgpt-on-wechat
      • chatgpt GPT Academic
      • RikkaHub
      • Coze Workflow Configuration Guide
    • Help Center
      • Thinking Model Response Format
      • HTTP Status Codes and Their Meanings
      • Image Upload API
    • Decisions API (TypeSafe JEV)
      • Create Decisions
        POST
    • 3D Generation (Hunyuan hy-3d)
      • Query 3D Generation Task
      • Submit 3D Generation Task
    • File Upload
      • Temporary Images
        • Temporary Base64 Image Upload
        • Get Presigned Direct Upload URL (recommended for large files)
        • Re-host Remote Image as Temporary URL
        • Temporary Local Image Upload
      • Playground
        • Playground Image Upload
      • Deprecated
        • Legacy Upload Endpoint (Unsupported)
  • Unified Video API
    • Unified Video
  1. Decisions API (TypeSafe JEV)

Create Decisions

默认 API 入口 api.crazyrouter.com
https://api.crazyrouter.com
默认 API 入口 api.crazyrouter.com
https://api.crazyrouter.com
POST
/api/alpha/decisions
OpenRouter Decisions API compatible endpoint for the typesafe/jev-* family of decision models (e.g. jev-1.13): given state (the state/text to be judged) and questions (a number of noul / choice / score type questions), the model returns structured answers. The request body is passed through byte-for-byte in the native OpenRouter format; only model is rewritten according to the channel mapping.
Constraints: model, state, and questions are required; each question needs type (noul | choice | score) and instructions, and choice/score also require non-empty criteria; streaming is not supported (stream must be omitted or false). jev models can only be called on this endpoint; on /v1/chat/completions they return 400 unsupported_endpoint_for_model.
Billing is based on the response usage.input_tokens / output_tokens multiplied by this site's prices; usage.cost in the response is the provider's cost field, not the amount charged by this site.
Source of truth in code: router/relay-router.go (decisionsRouter.POST("/decisions")), relay/decisions_handler.go, dto/decisions.go, docs/decisions-api.md.
Model name: the public model name on this site is jev-1.13 (jev-1.13 is also accepted). The model field in the response is the full version name returned by the upstream.

Last verified/modified: 2026-09-28 (API Base URL unified to https://api.crazyrouter.com; example model names follow what GET /v1/models returns); tested: stream=true / missing questions both return 400 invalid_request_params; jev models on /v1/chat/completions return 400 unsupported_endpoint_for_model; api.crazyrouter.com now allows this path

Request

Authorization
Provide your bearer token in the
Authorization
header when making requests to protected resources.
Example:
Authorization: Bearer ********************
Header Params

Body Params application/jsonRequired

Examples

Responses

🟢200成功
application/json
Bodyapplication/json

🟠400参数错误 / 不支持流式
🔴502上游响应无效
Request Request Example
Shell
JavaScript
Java
Swift
curl --location 'https://api.crazyrouter.com/api/alpha/decisions' \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
  "model": "jev-1.13",
  "state": "A customer requests a refund after being charged twice for the same order.",
  "questions": {
    "action": {
      "type": "choice",
      "instructions": "Choose the best next action for the support agent.",
      "criteria": {
        "refund": "Refund the duplicate charge",
        "deny": "Deny the request",
        "escalate": "Escalate to a human"
      }
    },
    "urgency": {
      "type": "score",
      "instructions": "Rate urgency from 1 (low) to 5 (high).",
      "criteria": [
        "1",
        "2",
        "3",
        "4",
        "5"
      ]
    }
  }
}'
Response Response Example
200 - 成功示例
{
  "id": "gen-dec-01hx",
  "model": "typesafe/jev-1.13-20260917",
  "provider": "TypeSafe",
  "answers": {
    "action": {
      "type": "choice",
      "choice": "refund",
      "confidence": 0.97
    },
    "urgency": {
      "type": "score",
      "score": 4,
      "confidence": 0.88
    }
  },
  "usage": {
    "input_tokens": 321,
    "output_tokens": 31,
    "cost": 1.3482e-05
  }
}
Modified at 2026-09-28 14:59:30
Previous
Image Upload API
Next
Query 3D Generation Task
Built with