Skip to content

Latest commit

 

History

History
40 lines (31 loc) · 1.7 KB

File metadata and controls

40 lines (31 loc) · 1.7 KB

SiftMarks Roadmap

SiftMarks is currently focused on becoming the best local-first repair and search layer for browser bookmarks.

Now

  • Local Next.js dashboard and API.
  • Chrome extension import from the browser bookmark tree.
  • Chrome sync-back for accepted cleanup suggestions.
  • SQLite storage under ~/.siftmarks by default.
  • Keyword search with SQLite FTS.
  • Memory-mode search that uses embeddings when available.
  • Rule-based duplicate, vague-title, and broken-link cleanup suggestions.
  • Optional AI summaries, tags, embeddings, title suggestions, rescue suggestions, and taxonomy moves.
  • MCP server for AI clients.
  • CLI for init, import, stats, doctor, search, rescue, export, index, and MCP startup.

Next

  • Make semantic search easier to set up and easier to understand in the UI.
  • Add stronger tests around import, duplicate detection, rescue suggestions, and sync plans.
  • Improve folder policy controls so users can define the shape of a cleaned bookmark bar.
  • Add richer dry-run and diff views before Chrome sync-back.
  • Document Chrome extension setup for common Chromium-based browsers.
  • Improve MCP result formatting for agent workflows.

Later

  • Firefox import and possible sync-back support.
  • Optional local full-page archives for selected bookmarks.
  • Better broken-link checking controls and retry policies.
  • Portable export/import bundles for moving a SiftMarks library between machines.
  • More local-model guidance for Ollama users.

Non-Goals For Now

  • SiftMarks is not a hosted bookmark sync service.
  • SiftMarks is not a replacement browser UI.
  • SiftMarks does not silently reorganize bookmarks in the background.
  • SiftMarks does not require an AI provider for the core local workflow.