Skip to content

JARVIS RD Assistant

A self-hosted research workspace for discovering, organizing, and interrogating scientific literature. JARVIS links generated claims to retrieved source text, combines literature work with spaced-repetition learning and project tasks, and can run model inference on infrastructure you control with Ollama.

If you configure a cloud model through LiteLLM, relevant prompts and source excerpts are sent to that provider. Generated outputs can still be incomplete or wrong; read Methods and limitations before relying on verification badges, consensus views, or extracted evidence.

Start here

  • Quick start — the one-screen happy path: install, create an admin, and analyze your first paper.
  • Install JARVIS — the shortest supported path from a new host to the finish-setup link.
  • First sign-in and setup — create the first administrator, invite people, and add a passkey at the final family address.
  • Read the User Guide — research workflows and every user-facing area of the application.

Explore