Skip to main content
Transluce’s platform provides the unified technology stack for developing AI evaluations and analyses. We enable teams of scientists to collaborate with agents at remarkable speed and efficiency, without losing understanding of scientific results. The platform is built upon a simple philosophy: we provide scientists full control and visibility into core experimental design — including its logic, data, and visualizations — then solve the remaining generic but critical operational problems: reproducibility; trustworthy and verifiable provenance; performant, reliable, and scalable execution; artifact storage and versioning; sharing and collaboration. The Transluce platform layers surrounding your experiment: Studio + Hub, behaviors, and fxtr. The Transluce platform layers surrounding your experiment: Studio + Hub, behaviors, and fxtr. There are three layers to the stack:
  1. fxtr is a modular computation and data management framework, purpose-built for AI science
  2. behaviors provides a library of primitives commonly used in AI behavior evaluations
  3. The studio+hub (coming soon) is a managed service that helps you store, execute, view, and share experiments
To get started, check out our documentation:

fxtr

behaviors

studio+hub

Coming soon.