Models & routing

Rescriptum supports 300+ language models through OpenRouter. You can manually select a model or let the auto-router pick the best one for each task.

Model picker

The model picker organizes models by provider (Anthropic, OpenAI, Google, DeepSeek, etc.) with collapsible sections. Each model shows pricing, context length, and a free badge where applicable.

Auto mode

Enable Auto mode to let Rescriptum select the optimal model based on your preferences: prioritize speed, cost, quality, or open-source models. Preferences are configurable in the Auto tab of the model picker.

Token treasury

Every API call is tracked in the Token Treasury with session, daily, and lifetime analytics. Set budget limits and get warnings before hitting them.