Features
Single API key for 300+ AI models
OpenAI-compatible endpoint — drop-in replacement
Intelligent cost-optimized routing across models
Automatic fallback to alternative models on failure
Response caching to reduce latency and cost
Per-request rate limiting and throttling
Usage analytics with cost breakdown per model
Custom model selection rules and priorities
Real-time model benchmarking
Vision model support (Gemini 2.5, GPT-4.1, Qwen3 VL)
Pay provider cost + fixed per-request fee
Load balancing across providers
Detailed logging and observability
API access only (no visual interface)
Supports models from OpenAI, Anthropic, Google, Meta, Mistral, Cohere, AI21, Together, Groq
Unified API for 400+ models across 70+ providers
Automatic failover routing when a provider goes down
Model Fusion: combine budget models to rival frontier performance
Subagent delegation: frontier models hand off tasks to cheaper workers
Speech/transcription APIs (May 2026)
Unified Image API with capability discovery across 30+ models (June 2026)
OpenAI SDK compatible out of the box
Server-side presets to survive model deprecation
Edge-deployed for minimal latency
Budget controls and spend limits
Credit-based pricing with no subscriptions
Model routing visualization for latency and cost
Private model support for enterprise deployments
Data policy-based routing and enforcement
MCP server for coding agents (June 2026)
