LocalDocu

LocalDocu AI: in progress for the desktop app

LocalDocu AI (in progress) is an open-weight, on-device model for the desktop app that plans document tasks from your intent and metadata — no cloud upload.

LocalDocu AI is in active development for the desktop app: an open-weight, local-device-friendly model that understands user intent and LocalDocu’s document-management capabilities, then runs complex multi-step work through those local engines. LocalDocu AI is Off in the current web release. We aim to release soon and welcome contributions from the open community.

What LocalDocu AI is designed to do

  • Run from the LocalDocu desktop application (not as a cloud AI document pipeline)
  • Use document metadata and your request — not document contents — to plan work
  • Drive LocalDocu’s existing document-management and PDF/Word operations for complex tasks
  • Use an open-weight model suited to capable local devices, chosen to preserve quality and performance rather than trade them away

Intended model (WIP)

  1. You describe what you want in plain language in the desktop app
  2. LocalDocu AI interprets intent using LocalDocu capability knowledge plus document metadata (names, types, counts, structure signals — not file contents)
  3. A validated command plan is produced against the LocalDocu command registry
  4. Deterministic local engines execute the plan on your device

LocalDocu AI is work in progress and not active in the web workspace yet. Document contents are not sent to a remote AI for processing in the current release. Join us on GitHub if you want to help ship it.