Skip to main content

​
Model Configuration

Myrm uses LiteLLM for unified access to 100+ models from any provider.

​
Adding API Keys

Navigate to Settings > Models or set environment variables: 
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
DEEPSEEK_API_KEY=...
GOOGLE_API_KEY=...

​
Smart Routing

The complexity router automatically selects the optimal model based on task requirements:
  • Simple queries use lightweight models
  • Complex reasoning escalates to frontier models
  • Cost and latency are optimized per request

​
Key Rotation

4-strategy key rotation maximizes throughput and availability across multiple API keys per provider.

​
Privacy Routing

Myrm’s Privacy Routing automatically selects cloud or local models based on data sensitivity — no manual switching required:
SensitivityRoutingData Handling
S1 — PublicCloud modelDirect to cloud
S2 — InternalCloud (after PII redaction) or localAuto-redaction or local routing
S3 — ConfidentialLocal model onlyData never leaves your machine
Privacy Routing wraps the model behind a standard interface. Agents, middlewares, and the execution loop are completely unaware of the routing — they interact with a normal model. Combined with Ollama, LM Studio, or vLLM for local backends, this enables fully air-gapped operation for sensitive workloads.

​
Fault Tolerance

The 14-layer error recovery system handles failures automatically:
  • Rate limit errors (4-strategy key rotation + credential pool)
  • Provider outages (Circuit Breaker with 3-tier cooldown + fallback presets)
  • Stream interruptions (token-level precise resume)
  • Response truncation (progressive output budget boost 2x → 3x → 4x)
  • Oversized images (automatic re-encoding and compression)
  • Model thinking mode errors (automatic mode adjustment and retry)
  • Empty responses (parameter adjustment and retry)
  • Iteration limits (grace-call summary — users never see a blank response)
See Error Recovery for the full 14-layer architecture.