Repolode
Submit
byenzyme

byenzyme/enzyme

Local-first indexing tool that reads your Markdown notes' folders, tags and links, then lets Claude Code or Codex search them by meaning.

View on GitHubWebsite
Stars
135
Forks
6
Open issues
1
Last commit
12 days ago

What is it?

Enzyme indexes a Markdown knowledge base so a coding agent such as Claude Code or Codex can find related notes written over months and cite them. It uses the folders, tags, links and dates already in your notes to decide what to read, rather than requiring you to restructure them first.

It works in two steps. A settings file (an .enzyme program) defines which folders, tags or links to read, and Enzyme samples excerpts across time from each one to generate questions about decisions and open issues, called catalysts. The index then ranks notes by both keyword wording and meaning, and each search result shows the catalyst that connects it to the query. Search runs on your machine; catalyst generation can use a hosted service, an OpenAI-compatible key, a local model, or be skipped with --llm none.

It also supports Apple Notes and Are.na through presets, and experimentally indexes SQLite tables such as message archives. Installation is a shell script with a CLI, and setup can be driven by an agent through an installed skill.

Key features

  • 01Builds a local search index over Markdown notes without moving or editing them.
  • 02Ranks results by combined keyword and semantic matching.
  • 03Generates per-topic questions (catalysts) to map recurring themes in notes.
  • 04Supports Apple Notes and Are.na presets, plus experimental SQLite sources.
  • 05Works with Claude Code, Codex, Hermes and OpenClaw via an install command.
  • 06Runs without a model key when catalyst generation is turned off.

Use cases

  • Searching an Obsidian vault for how a project or decision evolved over time.
  • Giving a coding agent access to meeting notes, journals or ADRs as context.
  • Finding recurring themes across dated notes about a person or relationship.
  • Keeping a private, local-only search over personal notes.

Quick start

  1. 1

    Run the official install script from the README

  2. 2

    Go to your Markdown notes folder

    cd /path/to/your/notes
  3. 3

    Install Enzyme for your agent (Claude Code, Codex, etc.)

    enzyme install claude

Things to know

  • Catalyst generation needs an API key or the hosted service; without one, indexing and local search still work but catalysts are skipped.
  • Message archive support from SQLite tables is marked experimental.
  • Setup can use Enzyme's online service to suggest settings, though you can review them before the index is built.

In collections

Claude Code tools, skills & pluginsPrivacy & local-first toolsStatic site generators & documentation tools

Topics

#knowledge-management#markdown#semantic-search#local-first#obsidian#claude-code#pkm

FAQ

What is enzyme?
Local-first indexing tool that reads your Markdown notes' folders, tags and links, then lets Claude Code or Codex search them by meaning. Enzyme indexes a Markdown knowledge base so a coding agent such as Claude Code or Codex can find related notes written over months and cite them. It uses the folders, tags, links and dates already in your notes to decide what to read, rather than requiring you to restructure them first.
Is enzyme free to use? What license does it have?
enzyme is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is enzyme still maintained?
Actively maintained. The last commit to enzyme was on Oct 26, 2026.
What can I use enzyme for?
Searching an Obsidian vault for how a project or decision evolved over time. Giving a coding agent access to meeting notes, journals or ADRs as context. Finding recurring themes across dated notes about a person or relationship. Keeping a private, local-only search over personal notes.
How do I install enzyme?
Run the official install script from the README Go to your Markdown notes folder: cd /path/to/your/notes Install Enzyme for your agent (Claude Code, Codex, etc.): enzyme install claude
Where can I get enzyme?
The source code and installation instructions are on GitHub: https://github.com/byenzyme/enzyme

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects