Model coverage
OpenRouter's catalogue is very broad, including many open-weight and niche models, and it can route a request across several hosts of the same model. The LazySusan API covers the leading chat models (GPT-6, Claude, Gemini, Grok, DeepSeek, Perplexity Sonar) plus image models (Nano Banana, GPT Image, FLUX, Ideogram, Recraft, Grok Imagine, Luma Photon) and video models (Veo 3.1, Runway, Luma Ray, Hailuo, Vidu).
edge: OpenRouter
Image and video
With LazySusan, starting a Veo 3.1 or Runway clip uses the same key and SDK as a chat call, and failed videos are refunded in full. OpenRouter is built around language models; if you need video generation, check what its catalogue covers today.
edge: LazySusan API
Routing and fallbacks
OpenRouter's provider routing, with fallbacks between hosts, is a genuine strength for uptime-sensitive text workloads. LazySusan calls each model at its own provider and returns a clear error rather than switching models behind your back.
edge: OpenRouter
Team controls
LazySusan has roles (owner, admin, developer, viewer), invites by email, keys per person, project and environment, monthly budgets and model allowlists on any key or teammate, key expiry, suspension, an audit log and email alerts for balance, spend, key budgets and error spikes. Requests outside a budget or allowlist are refused before they run.
edge: LazySusan API
Usage analytics
Every LazySusan response includes the usage and credits charged, and the console breaks spend down by person, key, project, environment, model and provider, with a searchable request log, up to 16 metadata tags per request and CSV export.
edge: LazySusan API
SDKs and integration
LazySusan ships its own SDKs for JavaScript/TypeScript and Python, or plain HTTPS from any language; every model takes the same request shape, so switching is a one-string change. OpenRouter is supported by a large ecosystem of tools and frameworks.
edge: Tie