Features
Adaptive routing with online learning from real traffic
Automatic failover under 50ms across 200+ models
Zero token markup – pass-through billing at provider rates
Per-workspace routing objective: cheapest, quality, balanced, adaptive
Full structured logs for cost, model, latency, and failure reasons
Guardrails and agent firewall with PII shield and content policies
OpenAI-compatible endpoint – drop-in replacement
OrcaRouter MCP server for agent framework integration
Prompt versioning with A/B splits and one-click rollback
Prompt caching across 5-minute and 1-hour windows
RouterArena-leading accuracy at 75.5% (June 2026)
Routing DSL for custom model composition (new June 2026)
Model fusion for combining model outputs (new June 2026)
Compliance enforcement and audit reporting (Team and up)
Risk-scored agent firewall with anomaly detection
Local model execution for offline inference
Connect online APIs: OpenAI (ChatGPT), Anthropic (Claude), Google (Gemini), Meta (Llama), Mistral AI, Alibaba (Qwen),
123+ HuggingFace models available
Memory feature (coming soon) for context persistence
Minimalist UI with dark mode
Open source under MIT license on GitHub
Cross-platform desktop app (Mac, Windows, Linux)
Version 0.8.3 with bug fixes and performance improvements