Score content in real time

API Integration

Evaluate any content against your preference profiles through a REST API designed for production workloads. Sub-100ms p99 latency, batch endpoints, webhooks, and SDKs for every major language.

Key Capabilities

Sub-100ms Latency

Preference scoring runs on optimized inference infrastructure. Median response times under 20ms with p99 under 100ms, suitable for real-time content pipelines and user-facing applications.

Batch Processing

Submit up to 10,000 frames in a single batch request. Results stream back as they complete, with automatic retry and partial-failure handling built in.

Webhooks & Events

Subscribe to lifecycle events such as annotation completion, profile updates, and quality alerts. Webhooks deliver signed payloads with automatic retry and dead-letter queuing.

SDKs & Libraries

Official client libraries for Python, TypeScript, Go, and Rust. Each SDK provides typed models, automatic pagination, retry logic, and streaming support out of the box.

Usage

Score a frame against a profile
curl -X POST https://api.commandagi.com/v1/score \
  -H "Authorization: Bearer $COMMANDAGI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "profile_id": "prof_main",
    "frames": [
      { "id": "frame_001", "url": "https://cdn.example.com/img1.png" },
      { "id": "frame_002", "url": "https://cdn.example.com/img2.png" }
    ],
    "dimensions": ["aesthetics", "usability"]
  }'

# Response
# {
#   "scores": [
#     { "frame_id": "frame_001", "aesthetics": 0.92, "usability": 0.78 },
#     { "frame_id": "frame_002", "aesthetics": 0.65, "usability": 0.91 }
#   ],
#   "latency_ms": 18
# }

Ready to get started?

Create your first preference profile in minutes.