Wenlan
Wenlan is a source-backed AI knowledge base and LLM wiki for AI work: organize documents and decisions into pages you can find, inspect, and review.
翻訳待ち
Structured Augmented Knowledge Orchestrator — an open-source knowledge system for organizing projects, decisions, context and AI-assisted workflows. Plain Markdown, a small CLI, and provider-neutral r
翻訳待ち
生産性Python · ShellGitHub @Sako404オンライン所有者未確認
Have you tried it? One tap is enough. No sign-in needed.
SAKO Brain — Structured Augmented Knowledge Orchestrator. Open-source, local-first knowledge system for organizing projects, decisions, context and AI-assisted workflows. Markdown is the source of truth.
翻訳待ち
Loading the full README. The saved excerpt is shown below.
Structured Augmented Knowledge Orchestrator —
Structured knowledge, projects, decisions, timeline and context;
Augmented by AI, tooling, indexing and automation;
Knowledge and context at the core; an
Orchestrator that coordinates knowledge, rules, context and the tools and agents that use them.Read the full README ↗Structured Augmented Knowledge Orchestrator — Structured knowledge, projects, decisions, timeline and context; Augmented by AI, tooling, indexing and automation; Knowledge and context at the core; an Orchestrator that coordinates knowledge, rules, context and the tools and agents that use them.
An open-source knowledge system for organizing projects, decisions, context and AI-assisted workflows — a local-first personal knowledge and project-management CLI. Your notes are plain Markdown files with YAML frontmatter, in directories you can read, edit and back up with anything. The search index, integrity manifests and logs are a rebuildable cache that lives outside your notes — so the vault stays Markdown and configuration, and nothing else.
It works from a plain text editor with no AI, no account and no network.
Current release: 0.24.1. Pre-1.0 means the command line and configuration format may still change; your notes will not — they are text.
Notes accumulate faster than structure does. Most tools answer this by owning your data: a database, a sync service, an account. That trade buys convenience and costs control — you cannot grep it, you cannot diff it, and you cannot take it somewhere else without an export feature someone else maintains.
Sako Brain takes the other side. The files are the system. Everything the program adds — full-text search, link and integrity checking, project records, decision records, a timeline, session handoffs — is derived from those files and can be thrown away and rebuilt with one command. If the program disappears tomorrow, you still have a directory of Markdown.
It also assumes AI agents will read and write your notes, and treats that as a
client concern rather than a feature. brain agents-doc --write generates an
AGENTS.md describing the vault's data model and rules in a provider-neutral
way, which any agent — or person — can read. It is opt-in: nothing about
the vault assumes an agent is involved, and brain init does not create one.
Being explicit, because these are reasonable things to expect and none of them is true:
brain doctor reports it if missing. It is a property of
your Python's sqlite3 module and cannot be installed by pip.| Linux | Supported and tested. |
| macOS | Expected to work but not tested. Reports welcome. |
| Windows | Not supported. File permissions rely on POSIX mode bits. |
| Scheduled backups | Linux with systemd only. |
None is a Python dependency, and pip will never install them. Each enables one feature; without it, that feature says so and everything else works.
| Tool | Enables |
|---|---|
git | brain git — local version history for your vault |
restic (and rclone for remote destinations) | brain backup — encrypted, versioned backups |
systemd | scheduled backup and maintenance timers |
Sako Brain is not on PyPI. Install the wheel from the latest release:
pipx install ./sako_brain-0.24.1-py3-none-any.whl
Each release lists the SHA-256 of its artefacts, so you can check what you downloaded:
sha256sum ./sako_brain-0.14.4-py3-none-any.whl
Or build from source:
git clone https://github.com/Sako404/sako-brain.git
cd sako-brain
python -m build
pipx install ./dist/sako_brain-0.14.4-py3-none-any.whl
pipx puts brain on your PATH and keeps it isolated. A plain virtualenv
works too:
python -m venv ~/.venvs/brain
~/.venvs/brain/bin/pip install ./dist/sako_brain-0.14.4-py3-none-any.whl
Either way, exactly one dependency is installed: PyYAML.
pipx install --force ./dist/sako_brain-0.14.4-py3-none-any.whl
pipx uninstall sako-brain
Uninstalling removes the program. It never touches a vault, its runtime state or its backups.
brain init ~/brain
Creates the vault, its directories and a commented configuration file, then
runs brain doctor and prints the result. It does not create an
AGENTS.md — see below.
It never overwrites anything. Run it again on an existing vault and it refuses;
with --force it fills in only what is missing and still leaves your
configuration and notes untouched.
Useful flags: --git also initialises version history. --no-agents is
accepted but does nothing — it is kept only so scripts written against 0.1.0
keep working.
brain agents-doc --write
Writes AGENTS.md at the vault root, rendered from that vault's own
configuration, describing the data model and rules in a provider-neutral way.
Run it whenever you want one; nothing else needs it, and a vault without one is
complete and passes brain doctor.
brain init --demo /tmp/example-brain
brain --vault /tmp/example-brain index
brain --vault /tmp/example-brain search observatory
Creates a small synthetic vault — fictional content, a deliberately different
directory layout, a custom note type — so you can see the shape of the thing
before committing to one. Safe to delete. Do not use --demo for a vault you
intend to keep.
This is the beginning of the README. Read the rest at the original. Read the full README ↗