Enzyme CLI
A CLI that builds a local notes index with catalyst questions so the agent finds material faster than plain grep search
- 145 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Catalyst generation sends note excerpts to the chosen model: hosted or local.
- Installing through `curl | bash` runs a script from the internet.
Install
- Works in
Manual install
curl -fsSL https://raw.githubusercontent.com/byenzyme/enzyme/main/install.sh | bashInstall for manual setup, then `cd /path/to/your/notes`.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Catalyst generation sends note excerpts to the chosen model: hosted or local.
- Installing through `curl | bash` runs a script from the internet.
This is third-party code. Review the repository files before installing.
What it does
Enzyme compiles Markdown notes and other knowledge sources into a local index. For each theme it writes questions ahead of time, called catalysts, and embeds them. At query time each question finds its nearest documents, so full summaries need not be written. During setup the CLI reads the vault and existing agent instructions and drafts an `.enzyme` program: reading profiles, a question budget, sources and folders. `enzyme petri` shows a map of themes, and `enzyme catalyze` searches by meaning and wording together. Obsidian, Apple Notes and Are.na are supported, with SQLite message tables as an experiment.
Who it is for. People with a growing notes vault: Obsidian, meeting notes, agent memory in Markdown, with questions that span months of entries.
Good fit when
- A project decision rests on two years of entries and they must be collected with source links.
- The agent has to find something that runs through many notes: a theme, a name, a recurring issue.
- An Obsidian vault grows daily and grep returns too much or nothing.
- The agent's context must stay small without losing links between old and new entries.
Not a fit when
- Exact search by titles and phrases is needed: the authors keep grep for that.
- Notes cannot go to external models: without a local model, catalysts need a key or a service.
- The vault holds few entries: the index pays off on growing bases.
Example request
Set up Enzyme for the open notes folder, show what to include in the index, then answer a question from the notes with source links.Limitations
Catalyst generation and the selection step of `enzyme compile` need an OpenAI-compatible key or a local model. By default `enzyme init` uses Enzyme's hosted bootstrap without your key. The first vault on a machine indexes without login, while refresh, publishing and extra vaults need `enzyme login`. Without keys, indexing and local search work but catalysts are not generated. Installation goes through `curl | bash`, so read the script before running it.
How to disable. Delete the `enzyme` binary, the agent instructions added by `enzyme install` and the `.enzyme` index folder in the vault.
FAQ
How do I connect Enzyme to Claude Code?
After installation run `enzyme install claude` in the notes folder. For other clients the README lists `codex`, `hermes` and `openclaw`.
What is a catalyst?
A question Enzyme wrote ahead of time from your notes and embedded in the index. At query time search runs through the nearest questions and each result shows the catalyst that connected it.
Is an API key required?
Only for catalysts and the selection step of `enzyme compile`. You can use a local model through `--llm env` with `OPENAI_BASE_URL` or turn generation off with `--llm none`.
Does Enzyme change my notes?
No. The setup prompt keeps notes in place unchanged, and cleanup, model downloads and recurring tasks are cleared separately with you.
Which reading profiles exist?
Seven: `operational`, `decision_trace`, `resonance_trace`, `reflective`, `relational`, `preference_evidence` and the default `tension_trace`. The profile is picked by note type, such as decisions or journals.
Discussion
No questions or reviews yet.
Related
CLI toolsObsidian Hybrid Search
obsidian-hybrid-search
Local Obsidian vault search: agents find notes by meaning, words, tags and links through MCP, CLI or a plugin.
Compass is an Obsidian vault template that takes journal questions and habits from one configuration file for its dashboards.
A Claude Code plugin that builds a knowledge system of markdown files from a conversation with you: folder structure, note templates, a processing pipeline and
Get笔记 CLI (getnote)
Get笔记 CLI
The getnote CLI saves links and text to Get笔记 (Dedao Brain), searches notes and knowledge bases and installs five agent skills.
An Obsidian vault with skills for work meetings: Gmail transcripts, participants, follow-up tasks and weekly recaps.
A plugin and skill set that turns an Obsidian vault into a knowledge base with sources, links and grounded answers
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Enzyme (this card) | no | 145 | ||
| obsidian-second-brain | no | 5,194 | obsidian-second-brain keeps agent memory in Obsidian; Enzyme searches existing notes through catalyst questions. | |
| knowledge-rag | no | not required | 301 | knowledge-rag turns a document folder into a base with hybrid search; Enzyme adds a question layer and accounts for time, tags and note links. |
README in short
The README explains the catalyst principle, an agent-driven setup prompt, manual setup with `compile`, `init` and `doctor`, Apple Notes and Are.na sources, search with `catalyze` and API key rules.