Hister v0.22.0

Hister is a free and open-source personal search engine that indexes the contents of web pages and local files so they can be searched later. It runs on infrastructure you control, provides a web interface and terminal search, and can also connect to AI assistants through MCP.

Features

Hister performs full-text indexing rather than searching only page titles and URLs. It can index pages visited through its Firefox and Chrome browser extensions and can also index local directories and imported browser history.

Its search system supports phrases, wildcards, field filters, negation, aliases, and result priorities. This provides more control than a basic keyword search when working with a large personal archive.

Hister also supports optional semantic search through a configurable embeddings endpoint. This allows searches based on meaning rather than exact words, although using a remote embeddings service means document text may be sent to that service.

The software can crawl websites recursively and import existing content. Recent releases have also added unified imports for browser history, local files, bookmarks, and archived content from services such as Linkwarden, Karakeep, and Shaarli.

Hister can be accessed through its web interface, terminal interface, TUI, command line, or an AI assistant using MCP. Multi-user support allows separate documents and search results to be maintained on a shared server.

Download Hister v0.22.0 - Software Mirrors

Hister v0.22.0 for Windows

nogon_0.22.0_windows_amd64.zip | 79.48 MB

nogon_0.22.0_windows_amd64.exe | 112.3 MB

Hister v0.22.0 for macOS

nogon_0.22.0_darwin_arm64 | 109.46 MB

nogon_0.22.0_darwin_amd64 | 111.6 MB

Hister v0.22.0 for Linux

nogon_0.22.0_linux_arm64 | 108.96 MB

nogon_0.22.0_linux_amd64 | 112.17 MB

Others Download related to Hister v0.22.0

nogon_0.22.0_checksums.txt | 560 B

Hister v0.22.0 Source Code

Hister v0.22.0 Source code (zip)

Hister v0.22.0 Source code (tar.gz)

Hister v0.22.0 Release Notes:

Hister Is Now Nogon

This is the first release under the Nogon name. The rename follows a trademark complaint about the previous name. Development continues with the same purpose: saving and searching your content on a server you control. Read Hister Is Now Nogon for the background and what the transition means for your installation. The project website is now nogon.eu, and new release downloads use nogon names. The application, terminal client, documentation, and browser extensions use the new branding. Existing extension installations receive the rename through their normal store updates and retain their settings. Nogon reuses existing Hister configuration and data. The rename itself requires no data move, import, or reindex. Follow the upgrade notes below before replacing an existing installation.

New Features

Settings in the Web Interface

The new Settings page edits application preferences, watched directories, indexing, crawler options, extractors, semantic search, hotkeys, and supported server and authentication settings. It shows where values come from, marks unsaved changes, and reports validation errors beside the affected fields. Settings supplied by the main configuration file, environment variables, or startup flags remain locked. Administrator access is required in multi user mode. Save and apply persists web selections without rewriting the main configuration file and briefly interrupts service while applying them. Sensitive changes require review. Connection and authentication changes apply temporarily; reconnect and confirm working access within two minutes. Cancellation, timeout, or restart restores the last confirmed configuration.

RSS and Atom Imports

nogon import rss FEED_URL imports the full pages linked from an RSS or Atom feed. Duplicate URLs are queued once, existing documents are skipped unless --force is supplied, and interrupted imports can resume with nogon index --job-id JOB_ID. The command runs once and exits; it does not subscribe to future updates or import text embedded in the feed.

Richer File Previews

Local file results open in fullscreen previews with a download action. Code previews include syntax highlighting and a copy button, and plain text keeps its line breaks and indentation. PDF previews show readable extracted text with page divisions when available. DOCX previews preserve paragraphs, standard headings, and basic text formatting. PDF and DOCX previews do not reproduce the original page layout or images.

Retained Data Cleanup

nogon purge removes selected old document versions, unpinned history, and completed crawl records before an explicit cutoff. For example, nogon purge --versions --history --crawl-jobs --before 2026-01-01 --dry previews the affected counts. Removing --dry asks for confirmation before deletion. The operation applies across all users and requires administrator access in multi user mode. Current searchable documents, pinned history, and unfinished crawls remain intact.

OIDC Administrator Groups

Configure admin_groups to manage administrator access through OIDC group membership, with groups_claim selecting the provider's group field. Each successful OIDC login grants or removes administrator access according to the current memberships. Keep an independent local recovery administrator in case the provider becomes unavailable or revokes access. Leaving admin_groups unset preserves manually managed roles.

Enhancements

  • Browser extension: the toolbar icon shows a red border when automatic
indexing is disabled. Manual indexing remains available. Orion extension connections are now accepted by the server.
  • Search results: document update dates are visible again.
  • Documentation: an expanded quickstart, heading links, code copy buttons,
and improved mobile navigation make setup and reference pages easier to use.
  • Dependencies: Go modules, npm packages, container bases, Nix inputs, and
GitHub Actions were updated.

Bug Fixes

  • Embedding requests better respect model context limits. Reindexing validates
the embedding service before discarding existing vectors and keeps keyword entries when individual documents cannot be embedded. Invalid vector dimensions and storage failures still stop the rebuild.
  • Bulk deletion uses the complete search configuration, and command line
deletion counts match the documents the operation can remove.
  • Opened result history preserves titles containing literal markup instead of
changing or hiding parts of them.
  • Docker starts correctly without a main configuration file. Directory
watching handles filesystem events more reliably, and invalid sensitive content patterns are rejected before use.
  • Browsers receive the correct OpenSearch description type. MCP requests
receive consistent protocol validation and structured errors for unknown methods, and MCP previews include stored PDF and DOCX previews even without stored HTML.
  • The browser extension recognizes case insensitive indexing rules beginning
with (?i) and avoids duplicate rule requests while pages load concurrently.
  • Numeric web settings remain editable after saving. Fresh OAuth or password
logins can confirm settings changes when proxy authentication is enabled.

Backward Compatibility Notes

The following changes require attention when upgrading from v0.21.0:
  • Downloaded executables: tagged and rolling releases provide Nogon
binaries only. Install the executable as nogon, or nogon.exe on Windows, and update scripts, shortcuts, service launch commands, and download URLs. Windows releases also provide a ZIP containing only nogon.exe. Package managers may retain Hister names during the transition; follow the installation guide.
  • Existing installations: stop the server and make a complete backup before
upgrading. Keep existing data paths, service identities, Docker volumes, and the Compose project name. Nogon configuration paths take precedence over legacy paths, so avoid creating a second configuration unintentionally. Check nogon config path and nogon config show with the service's usual flags and environment. If multiple directories contain stored data, select the intended one with app.directory or NOGON_DATA_DIR. Keep the same server URL and credentials in your clients.
  • Environment variables: prefer NOGON__*, NOGON_CONFIG,
NOGON_DATA_DIR, and NOGON_PORT for new configurations. Their Hister equivalents remain supported. A present Nogon configuration variable takes precedence over its legacy equivalent, even when empty. Service import credentials also accept NOGON_IMPORT_ while retaining HISTER_IMPORT_.
  • Configuration generation: nogon config create now emits commented
examples. Use --full if scripts require active YAML values. Existing explicit settings retain their meaning and lock the corresponding web controls; comment out or remove a setting to manage it through the web page.
  • Removed command aliases: replace create-config with config create,
and create-user, show-user, update-user, and delete-user with user create, user show, user update, and user delete. These previously deprecated aliases have been removed.
  • Custom integrations: use POST /api/add for document submission;
obsolete submission routes have been removed. MCP transport requirements are validated more strictly, including request content types and accepted response formats. Check custom clients against the current MCP request examples.
  • Existing file previews: import or submit PDFs and DOCX files again to
recover page divisions or formatting that was not stored previously. Use nogon reindex for watched local files. This is optional and separate from the project rename.
  • Recent searches: the browser's recent search hiding controls have been
removed. Previously hidden searches can reappear on the home page; hiding them did not delete their history from the server.

Performance and Compatibility

Hister is written in Go and can run locally or on infrastructure controlled by the user. It supports SQLite and PostgreSQL as database backends, with PostgreSQL support added in version 0.12.0.

The application is designed for continuous personal indexing. Once the server is running, the browser extension can automatically send visited page content to the configured Hister server for indexing.

Hister does not require a cloud service or account for a local personal setup. The project states that it has no telemetry or mandatory cloud synchronization. Indexed documents and search indexes remain on the Hister server.

The current release series also includes features for public search deployments, documentation datasets, and additional import sources. Version 0.17.0 introduced a unified import interface and support for importing content from Linkwarden, Karakeep, and Shaarli.

System Requirements

Hister provides prebuilt binaries for supported platforms, including Windows, Linux, and macOS. The Windows version can be started directly from PowerShell using the hister.exe binary.

For normal use:

  • Windows, Linux, or macOS

  • Firefox or Chrome if browser history indexing is required

  • Storage for indexed documents and search indexes

  • A local or server environment where Hister can remain running

For building from source:

  • Go 1.26

  • npm

  • C compiler for CGO dependencies

Hardware requirements are not specified as a fixed minimum because resource usage depends on the amount of content being indexed and the selected database and search configuration.

Pros and Cons

Pros

  • Free and open source

  • Full-text personal search

  • Indexes web pages and local files

  • Firefox and Chrome integration

  • Browser history import

  • Website crawler

  • Powerful search operators

  • Optional semantic search

  • Web interface

  • Terminal and TUI access

  • MCP support for AI assistants

  • Multi-user support

  • No telemetry by default

  • No mandatory cloud service

  • Supports SQLite and PostgreSQL

  • Docker and Nix installation options

  • AGPLv3 licensed

Cons

  • Requires running and maintaining your own server

  • Initial indexing can consume significant storage and processing resources

  • Semantic search requires an embeddings endpoint

  • Browser integration requires an extension

  • More complex to configure than conventional desktop search tools

  • The AGPLv3 license may require consideration for organizations modifying and redistributing the software

How to Install

Download the Hister binary for your operating system from the project's releases. On Windows, rename the executable to hister.exe if necessary and start it from PowerShell with:

.\hister.exe listen

On Linux or macOS, make the binary executable and start it with:

chmod +x hister
./hister listen

Then open the local Hister web interface at http://127.0.0.1:4433 and install the Firefox or Chrome browser extension. After visiting pages with the extension enabled, their content can be indexed and searched through Hister.

Hister can also be installed using Homebrew, Docker, or Nix. Existing browser history, local directories, and files can be imported after installation.

Final Verdict

Hister is a useful personal search engine for users who want to build their own searchable archive of web pages and files. Its full-text indexing, browser integration, local deployment, advanced queries, semantic search option, and MCP support make it more than a simple browser history search tool.

The main tradeoff is that Hister is self-hosted software. Users need to keep the server running and manage the storage and indexing environment themselves. For users who want a private search index under their own control, that approach provides considerably more control over their data.

Hister v0.22.0
Free
Software Informations:
Developer:

Operating System:
Windows / macOS / Linux
Date Added:
2026-11-19T15:07:22.366Z
Categories:

Post a Comment/Report Broken Link: