Hister: A private search engine for the pages you visit and the files you keep

Original Article ↗ Hacker News Discussion ↗

A new open-source tool called Hister turns your browsing history, bookmarks, local files, and even social media content into a private, full-text searchable index that lives on your own machines. Commenters compare it to past efforts like Google Desktop, ArchiveBox, Zotero, and various DIY setups, praising its ability to surface previously visited pages while debating trade-offs around local vs. cloud hosting, browser integrations, and long-term archival. The project’s author weighs in on planned features, multi-device support, legal pressure to rename the project, and potential integrations with local LLMs and other self-hosted services.

Overview & Purpose

  • Tool builds a private search engine over pages you visit and files you keep.
  • Focus is on personal knowledge retrieval, not general web search.
  • Works as an index of visited pages, bookmarks, browser history, local files, and crawled sites.

Features & Architecture

  • Stores extracted page content with offline previews, so results remain even if the source changes or disappears.
  • Supports full‑text and semantic search, has a web UI, CLI, and MCP endpoint for AI/assistant integrations.
  • Browser extension indexes what’s actually rendered in tabs, allowing it to capture content behind logins and on “closed” sites (Instagram, Discord, Twitter, YouTube transcripts via yt-dlp, etc.).
  • Can be used from multiple machines and in multi‑user setups; no mandatory cloud component.

Comparisons to Other Tools

  • Compared to ArchiveBox: this focuses on fast knowledge retrieval; ArchiveBox on long‑term preservation. Some users want a single shared archive, not two.
  • People mention similar or adjacent tools: linkwarden/linkding/Betula, Zotero, SingleFile, Google Desktop, old Chrome/Opera history search, recoll, and various homegrown history-indexers.
  • Some see it as a modern equivalent of powerful local search tools that browsers have since “nerfed.”

Privacy, Security & Hosting Concerns

  • Strong interest in keeping data local and avoiding dependence on cloud LLMs; others are comfortable fronting it with ChatGPT or local models as a “reasoning layer.”
  • Some are wary of running anything outside distro repositories or worry about security vulnerabilities; suggestions include containers, AppArmor, systemd hardening, and strict network segmentation.
  • A few prefer low‑tech alternatives like printing pages to PDF plus pdfgrep.

Real‑World Usage Reports

  • Several users report weeks–months of use and call it “invaluable” for finding forgotten links, indexing notes/Obsidian, and making social/web content searchable.
  • One report of severe resource issues when reverse‑proxied via Nginx; root cause unclear.

Limitations & Feature Requests

  • Desired features:
    • Notes/annotations on pages, bookmark‑only filters, visibility‑time thresholds (e.g., only index tabs viewed >4s), keeping old versions and diffs, email indexing, Android/Chrome support, more bookmark manager integrations.
  • Some question personal ROI of elaborate knowledge bases and prefer minimalism.

Naming & Trademark Dispute

  • Current name conflicts with an existing US trademark; maintainers expect to rename.
  • Thread contains many alternative name suggestions and debate over whether a rename is legally necessary.