Open source wikis you can self-host, tested

Wikis and knowledge bases you can run yourself. The list is ordered by whether each project came up when Argusic installed it fresh, with the recording of that attempt one click away.

Tested between and . Each row shows its own test date; a project can change after that day.

94 of 96 tested projects run. 7 more waiting for a test.

In short: 74 of the 96 tested projects started as-is on a fresh machine: book-to-skill, claude-obsidian, obsidian-mind, obsidian-second-brain, colanode, Raneto, karpathy-llm-wiki, and Matterwiki, and 66 more. 20 more started once a stand-in replaced a service they expect, such as a database: mm-wiki, trustgraph, claude-chat-exporter, OpenTagViewer, wiki, compound-memory, myPKA, and project-cairn, and 12 more. 2 could not be verified: OpenWiki and indie-wiki-buddy; the log shows where each one stopped.

Measured by Argusic on a fresh machine every time. Every number links to its evidence.

#projectverdictArgusic Scorelanguagestarstested on
1book-to-skillRuns100 / 100Python38,736

Turn any technical book PDF into a Claude Code skill, ready to study, reference, and use while you work.

What the test found: The book-to-skill package is installed as an editable pip package, passes ruff lint, all 627 tests pass, skill validation passes, and the CLI extracts text from documents producing structured JSON metadata and clean text output. 3 minutes.

2claude-obsidianRuns100 / 100Python16,691

Self-organizing AI second brain for Obsidian + Claude Code. Drop any source and Claude reads, links, and files it into one connected knowledge graph of plain...

What the test found: The 15 Codex skills installed successfully, and the full test suite, vault lifecycle, local BM25 retrieval, strict lint, and a reproducible 199-file release build all passed using only local resources. 5 minutes.

3obsidian-mindRuns100 / 100TypeScript5,555

A self-organizing Obsidian vault that gives AI coding agents persistent memory.

What the test found: 488 unit tests pass, QMD semantic search indexes 28 markdown files and returns ranked results, SessionStart hook runs cleanly without assertion errors or subprocess timeouts. 71 minutes.

4obsidian-second-brainRuns100 / 100Python5,422

Persistent memory for Claude Code, Codex and 6 other AI agents, stored as plain markdown in your Obsidian vault. New: Crew, one shared logbook so Claude Code...

What the test found: All 996 tests pass, 8 platform builds succeed, vault health and lint checks pass, install completes with 47 commands linked and SessionStart hook registered. 26 minutes.

5colanodeRuns100 / 100TypeScript5,216

Open-source and local-first Slack and Notion alternative that puts you in control of your data

What the test found: All packages build and type-check; server ES bundle compiles; web app builds and all 133 tests pass. 15 minutes.

6RanetoRuns100 / 100JavaScript2,906

Markdown powered Knowledgebase Wiki for Node.js

What the test found: Raneto builds and runs with the downloaded Node v26.13.0: the server listens on port 8080, serves pages with HTTP 200, renders Markdown content, and accepts admin login via /rn-login. 4 minutes.

7karpathy-llm-wikiRuns100 / 100Python2,789

Agent Skills-compatible LLM wiki for Claude Code, Cursor, and Codex. Build a Karpathy-style knowledge base from raw sources, citations, and linting.

What the test found: All 52 tests pass and the check_evidence.py script verifies wiki article fidelity against raw source material, detecting mismatched numbers, dates, and quotes. 3 minutes.

8MatterwikiRuns100 / 100JavaScript2,090

A simple and beautiful wiki for teams

What the test found: Matterwiki runs on Node 18 with webpack-built frontend (bundle.js), Sqlite3 database, JWT authentication, and all API endpoints (articles, topics, users, search, archives, setup, admin) are operational. 13 minutes.

9TidGi-DesktopRuns100 / 100TypeScript2,013

TidGi is an privacy-in-mind, automated, auto-git-backup, freely-deployed knowledge management Desktop note app, based on Tiddlywiki, with REST API for...

What the test found: TidGi desktop app dependencies installed, 100 test suites pass (677 tests), TiddlyWiki plugins compile, and the Electron app launches and initializes workspaces successfully. 17 minutes.

10OpenLoreRuns100 / 100Go1,713

A minimal, extensible, agent-native knowledge repository that keeps shared context current and inspectable

What the test found: OpenLore builds from source, all 23 test packages pass, the server binary starts on port 8080 (HTTP), 2222 (SSH), and 3000 (metrics), serving docs over HTTP, MCP, and a JSON API - verified by web requests returning 200 and the shell API returning document content correctly. 6 minutes.

11otterwikiRuns100 / 100Python1,587

A minimalistic wiki powered by python, markdown and git.

What the test found: All 783 tests pass on Python 3.12.3; the Otter Wiki app installs, starts, and serves pages correctly using a local git repository with no external dependencies. 5 minutes.

12llm-wikiRuns100 / 100Python1,565

LLM-compiled knowledge bases for any AI agent. Parallel multi-agent research, thesis-driven investigation, source ingestion, wiki compilation, querying, and...

What the test found: llm-wiki Codex plugin v0.25.0 installs from local checkout via bootstrap-codex-plugin.sh, all 15 structural and runtime test suites pass (0 failures), and the scripts/llm-wiki CLI (lint, archive, schema, retract, specialist, adapter, checkpoint) parses and executes correctly. 4 minutes.

13COG-second-brainRuns100 / 100Python1,512

Self-evolving second brain with 35 AI skills, 10 agents, and people CRM. Closed-loop harness: a V-model verification lifecycle where the worker never grades...

What the test found: COG v3.15.0 is installed and fully validated: all 36 Claude Code skills, 36 Antigravity skills, 7 Kiro powers, 7 Gemini commands are present and verified, the Agent Plugins mirror is in sync, the Python-based slop-gate scanner runs and detects AI-slop tells, the lane-classify and checkpoint harness scripts produce... 5 minutes.

14leafwikiRuns100 / 100Go1,240

LeafWiki - Self-hosted wiki. Single Go binary, SQLite, Markdown on disk. No external database required.

What the test found: LeafWiki compiles from source, starts on port 8080 with SQLite storage and JWT auth, health API returns 200 OK, admin login succeeds with cookie-based JWT tokens, page tree is readable, pages can be created/read/updated via the API, and the full Go test suite (48 packages) passes with zero failures. 12 minutes.

15many-notesRuns100 / 100PHP1,100

Markdown note-taking web application designed for simplicity

What the test found: PHP 8.4.23 static binary installed, Composer deps resolved, Node 26 frontend built, SQLite database migrated, and Pest test suite 179/179 passes verifying all Feature and Unit tests. 27 minutes.

16crowiRuns100 / 100TypeScript1,099

Crowi - The Markdown Wiki - Empower the team with sharing your knowledge

What the test found: All 24 workspace packages build successfully, the API test suite passes 232/242 suites (10 smoke tests skipped requiring Redis), and the API TypeScript compilation produces dist/app.js. 19 minutes.

17PoznoteRuns100 / 100JavaScript1,050

Powerful note-taking without the hassle.

What the test found: Poznote installs and runs with PHP 8.3.6 CLI SAPI: 652 test assertions pass, built-in web server serves api_health.php with 200 OK returning status ok, service poznote, version 6.145.0. 17 minutes.

18codealmanacRuns100 / 100TypeScript949

A codebase wiki for AI coding agents. Captures what the code can't say: decisions, flows, invariants, gotchas.

What the test found: codealmanac 0.4.7 is installed and functional: all 564 tests pass, ruff linting passes, the CLI reports version 0.4.7, doctor shows green across install/python/database/manual/index/health, and validate confirms the repo's own 71-page wiki is valid. 1 minute.

19projectmemRuns100 / 100Python878

Open-source coding agents memory. Records issues, attempts, fixes and decisions, then warns your agent before it repeats an approach that already failed...

What the test found: projectmem 0.3.3 is installed in a venv at /work/repo/.venv, all 279 tests pass, the pjm CLI works end-to-end (init, doctor, help), and the MCP server module loads without error. 1 minute.

20brain.mdRuns100 / 100JavaScript721

A persistent, file-based memory layer for coding agents, give Claude Code, Codex & others a project brain (durable decisions, requirements, constraints) via a...

What the test found: All 50 project tests pass. Every CLI subcommand runs correctly: init scaffolds BRAIN.md and brain directory, CRUD page operations work end-to-end, hooks install/uninstall cleanly, and wiring creates agent config files. Zero external dependencies or credentials required. 4 minutes.

21ngx-scannerRuns100 / 100TypeScript666

Angular QR code, Barcode, DataMatrix, scanner component using ZXing.

What the test found: All 6 Angular/Karma unit tests pass (1 for the scanner library component, 5 for the demo app), and the library builds to dist/zxing-scanner via ng-packagr. 10 minutes.

22Athena-PublicRuns100 / 100Python649

Turn stateless LLMs into a stateful, compounding AI agent. Local-first persistent memory, quantitative decision screens, and governance harness across Claude...

What the test found: Athena SDK v9.9.9 installed with all dependencies, CLI boots and runs health checks, full test suite passes (368/368) in an ephemeral container with no credentials required. 2 minutes.

23pi-llm-wikiRuns100 / 100TypeScript636

Self-maintaining, Obsidian-compatible knowledge base for pi, turn raw sources into an interlinked wiki that compounds. Native Open Knowledge Format (OKF) v0.2.

What the test found: Node 22/pnpm install, typecheck, lint and the full 73-file vitest suite (921 passed) succeed, and the built MCP server runs a real bootstrap-capture-ingest wiki cycle in /tmp/wiki-e2e without any API keys. 3 minutes.

24mkdocs-awesome-navRuns100 / 100Python628

A plugin for customizing the navigation structure of your MkDocs site.

What the test found: Project installs via uv sync, all 249 unit tests pass, and the MkDocs documentation site builds successfully with the awesome-nav plugin active. 4 minutes.

25TangentRuns100 / 100TypeScript589

The Tangent monorepo.

What the test found: npm ci installs all deps, npm run build succeeds across all 7 workspaces, 562 vitest unit/integration tests pass, 9 of 12 Playwright e2e Electron tests pass, and the Tangent app main process launches and initializes on the virtual display. 8 minutes.

26WikiDocsRuns100 / 100TypeScript534

📗 Just a databaseless markdown flat-file wiki

What the test found: Backend NestJS API boots and serves at port 4321 on http://[::1]:4321; all 30 backend e2e tests pass; document CRUD, tree, search, auth, accounts, settings, and guest endpoints all respond correctly with real datasets on disk. 7 minutes.

27opencrewRuns100 / 100Shell498

Openclaw多智能体协同系统 | Multi-Agent OS for Decision Makers, 基于 OpenClaw (Clawbot) + Slack,让 AI 团队各司其职、自主稳定迭代。

What the test found: OpenClaw 2026.10.5 gateway runs with 7 OpenCrew agents (main, cos, cto, builder, ko, ops, research) and real OpenRouter model inference works. 13 minutes.

28lineage-skillRuns100 / 100Python488

Distill your teacher into a source-backed Skill that teaches you until you can work on your own.

What the test found: bash install_as_skill.sh --target codex exits 0 and installs a complete skill to /home/runner/.codex/skills/lineage-skill whose scripts run; the 10-test suite passes, the text-only pipeline generates a validated mentor skill, and the micro-lesson runtime reaches phase complete. 4 minutes.

29ViNoteRuns100 / 100Python486

ViNote(视记AI)|DeepSeek Harness SDK 驱动的开源视频知识 Agent。以工具编排、SQLite 会话恢复和流式交互,串联 YouTube / Bilibili 检索、下载、转写与笔记生成,支持视频问答、知识卡片和思维导图。FastAPI + React,支持 OpenAI 兼容接口模型。

What the test found: ViNote v1.4.0 backend serves SPA on port 8999, SQLite initialized with 17 categories, all 50+ API endpoints respond, full test suite passes (214 backend + 18 frontend + media smoke). 15 minutes.

30okf-skillsRuns100 / 100Python456

The OKF toolkit for Claude Code, author, maintain, validate & visualize Open Knowledge Format bundles. Plugin, agent skills, and a GitHub Action.

What the test found: All test suites pass (208/208), both OKF bundles validate conformant under --strict, MCP server advertises 3 tools on stdio transport, visualizer generates cytoscape HTML graphs from bundles, and backfill events extractor runs against git history. 2 minutes.

31llm-wikiRuns100 / 100Python403

LLM-powered knowledge base from your Claude Code, Codex CLI, Copilot, Cursor & Gemini sessions. Karpathy's LLM Wiki pattern, implemented and shipped.

What the test found: llmwiki runs in a Python 3.12 venv: setup, sync (converts real Codex sessions), build (90-page HTML site plus llms.txt/graph.jsonld/manifest/sitemap/rss), serve (HTTP 200 on /), unit tests (2650 passed), and real-browser E2E (193 passed) all completed successfully; only index.md was not updated to list the new... 8 minutes.

32Wiki_MkDocsRuns100 / 100Python389

基于 MkDocs & Material theme 的个人知识库

What the test found: MkDocs 1.6.1 and MkDocs Material 9.7.7 are installed in .venv, 'mkdocs build -d public' completes with 1492 HTML pages, and the dev server answered HTTP 200 on http://127.0.0.1:8000/ serving the wiki. 4 minutes.

33project-butlerRuns100 / 100Shell377

Project memory system for AI coding assistants (Claude Code, Cursor, Codex): session logs, project wiki, rules, TODOs, and handoff.

What the test found: The skill is installed at ~/.claude/skills/project-butler and its only executable, scripts/check-update.sh, runs correctly in up-to-date, update-available, silenced, pull-ok, pull-failed, and offline-fallback modes; the repo has no test suite and no Claude Code CLI is available here to run the interactive... 2 minutes.

34linkitRuns100 / 100JavaScript356

AI bookmark manager for desktop, local-first, offline, read-it-later and PKM tool with AI summaries, semantic search, knowledge graph, and free self-hosted...

What the test found: The Homebrew cask update script's 8 unit tests pass on Node v18 after a one-line import.meta.dirname fix; the Go utility and Wails desktop app are not buildable in this environment. 2 minutes.

35mcp-logseqRuns100 / 100Python352

MCP server to interact with LogSeq via its Local HTTP API - enabling AI assistants like Claude to seamlessly read, write, and manage your LogSeq graph.

What the test found: mcp-logseq 1.10.0 installed, builds, and the full test suite of 720 unit and integration tests passes cleanly. 1 minute.

36stashRuns100 / 100Python345

Open-source infrastructure for agents that learn from experience. Capture production traces and turn lessons into reusable knowledge and skills.

What the test found: All four test suites pass (CLI 259, frontend 336, www 14, backend 184), both Next.js apps build, and ruff lint/format passes. 24 minutes.

37confluence-cliRuns100 / 100JavaScript330

A powerful command-line interface for Atlassian Confluence

What the test found: npm install succeeds, all 1232 tests pass, and the CLI binary prints its help output correctly. 1 minute.

38VeritasGraphRuns100 / 100Python326

VeritasGraph, open-source Knowledge Graph & GraphRAG framework on GitHub. Build multi-hop reasoning, ontology-aware retrieval, and verifiable attribution over...

What the test found: The VeritasGraph MCP server and Studio API install, launch, and pass 42 core tests with all JSON-RPC methods (initialize, tools/list, resources/list, resources/read, ping) responding correctly and the health endpoint returning 200 OK. 16 minutes.

39claude-canvasRuns100 / 100Python307

AI-orchestrated visual production for Obsidian Canvas. Create presentations, flowcharts, mood boards, knowledge graphs, and galleries with intelligent layout...

What the test found: Python test suite passes 103/103, all 12 templates instantiate and validate correctly, all 6 layout algorithms produce valid canvases, validation catches all error/warning cases correctly. 2 minutes.

40collabmdRuns100 / 100JavaScript292

Realtime collaboration for markdown folders, diagrams, and git-backed docs.

What the test found: CollabMD v0.1.49 builds, lints, and passes all 924 non-E2E tests (648 unit + 131 integration + 139 browser + 6 guardrails) under Node.js 26.0.0. The server starts on localhost:1234 and serves the app (HTTP 200 on / and /health endpoints) against a plain Markdown vault. 7 minutes.

41wikitRuns100 / 100JavaScript292

Wikipedia summaries from the command line

What the test found: Wikit v4.4.1 is installed with all dependencies, and successfully fetches Wikipedia summaries with flags for language, line wrapping, full articles, links, and disambiguation. 1 minute.

42markdown-confluenceRuns100 / 100TypeScript289

Publish your Markdown Files to Confluence

What the test found: pnpm install --frozen-lockfile succeeds, unit tests pass (606/607), static checks pass, all 6 packages build, integration test passes with real Mermaid rendering in Chromium and CLI fixture parity. 9 minutes.

43bkmrRuns100 / 100Rust269

Knowledge Management for Humans and Agents

What the test found: bkmr builds from source, all 384 tests pass, and the CLI works end-to-end (create-db, add, search, show with JSON output). 11 minutes.

44mnemeRuns100 / 100JavaScript252

The memory that dreams, long-term memory for AI agents. Readable files on your own machine, consolidated between sessions. Runs as a DeepSeek Harness plugin or...

What the test found: All 2587 tests pass on node 22.23.4 with the libc inference fix in build-runtime-manifest.mjs and runtime-manifest.json regenerated; src/ and lib/ synced; both prior failing tests (try 1777 libc filtering and 1778 manifest drift) now pass. 31 minutes.

45openNAMURuns100 / 100Go252

여러 기능이 있는 위키 엔진 (Wiki engine with multiple functions)

What the test found: openNAMU v4.3.6-beta.2 builds from source and runs its HTTP wiki server; registration, login, document creation with rendering, content search, history, and emergency-tool were all exercised successfully against a fresh SQLite database. 16 minutes.

46ai-fomo-skillsRuns100 / 100Python246

Personal superalignment skills for turning AI information overload into reusable knowledge, signals, and digests.

What the test found: The 3 skill directories are installed in ~/.codex/skills. The import_source.py script handles 6 source types: official-page, rss, github-repo, github-trending (all work without credentials), x-user (requires X_BEARER_TOKEN), and xiaoyuzhou-episode (scrapes public episode page). The xiaoyuzhou_account.py script... 21 minutes.

47keep-the-whyRuns100 / 100JavaScript240

The why layer of repo-native project memory: the reasoning behind a codebase as Markdown in the repo, versioned by Git, for coding agents and humans, so...

What the test found: keep-the-why-lint 0.20.0.0 and keep-the-why-dashboard 0.7.3 are installed and functional. The linter validates this project's context/ (0 errors, 0 warnings, schema 0.20.0). The dashboard serves its UI at localhost:8765 (HTTP 200), exports static pages, and both packages' test suites pass completely (116+61+65 = 242... 4 minutes.

48Pepperminty-WikiRuns100 / 100PHP208

A wiki in a box

What the test found: Pepperminty Wiki v0.25-dev built from source on PHP 8.3.29 in this container; the CLI version/help and search-rebuild commands run cleanly, and a fresh wiki installed via the HTTP first-run wizard served pages, handled login and page save, answered the search and status APIs, and exported a valid zip containing the... 10 minutes.

49Personal_External_BrainRuns100 / 100Python200

Not another ChatGPT wrapper. This is a local-first tool that turns your notes, web pages, PDFs and audio/video into a personal knowledge graph, running...

What the test found: Backend API serves 30 endpoints with seed graph data, all 131 automated tests pass, and the React frontend builds and starts. 4 minutes.

50booklemeRuns100 / 100Python174

读了么(BookLeme)· 把一本书编译成 Agent Skill, 双引擎 + 可插拔体裁插件 + 逐字忠实度回验

What the test found: BookLeme 1.0.0 installed with zero third-party dependencies, 34/34 unit tests pass, the CLI pipeline compiles a synthetic book into both a knowledge_pack and behavior_skill artifact with all validation gates passing, all JSON schemas validate, both built-in genre plugins pass validation, and both example skill... 1 minute.

51open-spacesRuns100 / 100JavaScript4

Free, open-source alternative to ChatGPT Spaces: a self-hostable AI workspace for shared pages and conversations.

What the test found: API server on port 8000 and Next.js client on port 3000 both serve HTTP 200 responses; the API returns/creates spaces, pages, messages, and auth records correctly over a local SQLite database. 10 minutes.

52tmsRuns97.1 / 100JavaScript389

基于频道模式的团队沟通协作+轻量级任务看板,支持mardown、富文本、在线表格和思维导图的团队博文wiki,i18n国际化翻译管理的响应式web开源团队协作系统。

What the test found: TMS Spring Boot app compiles and runs on port 8080 with MySQL 5.7.44 backend; 27 of 33 tests pass. 16 minutes.

53obsidian-wikiRuns96 / 100Python3,798

Framework for AI agents to build and maintain a digital brain through Obsidian wiki | Memory System for Agents

What the test found: obsidian-wiki package installs from source in a venv, setup writes config and provisions skills into 12 agent directories, 1029/1032 tests pass, doctor reports all-clear, the FastAPI/MCP server starts on demand and serves health, search, pages, memory, context-pack, staging, and status endpoints over HTTP. 8 minutes.

54TW5-TiddlyMapRuns96 / 100JavaScript914

Map drawing and topic visualization for your wiki

What the test found: TiddlyMap v0.17.21 builds with gulp --production (node v18.19.1), and serves as a plugin under TiddlyWiki 5.2.2 when both the compiled dist and a vis-network mock plugin are on TIDDLYWIKI_PLUGIN_PATH. Server returns 200 with full wiki content and plugin tiddlers are accessible via the REST API. 9 minutes.

55FridayOS1.0Runs96 / 100Python899

Local-first AI second brain & personal knowledge management: Six-Brain Markdown blueprint, Obsidian example vault, offline deployment and bilingual guides...

What the test found: All three repo-local Python scripts run successfully: the demo brain generator produces 129 wiki files matching the shipped zip content, the transcript parser converts 188 JSONL files to 179 markdown docs, and the star history script works offline. The demo zip (182 entries) passes CRC integrity checks and all media... 9 minutes.

56deepwiki-openRuns94.7 / 100Python18,473

Open Source DeepWiki: AI-Powered Wiki Generator for GitHub/Gitlab/Bitbucket Repositories. Join the discord: https://discord.gg/gMwThUMeme

What the test found: Backend FastAPI server runs on port 8001 with health, models, auth, repo, wiki, chat, codemap, and system endpoints all responding; frontend Next.js dev server compiles and serves on port 3000; 100 out of 105 tests pass (2 deselected for real API keys, 3 stale/invalid tests). 21 minutes.

57silverbulletRuns90 / 100TypeScript6,575

The malleable knowledge base for you and your team. Powered by Markdown and Lua.

What the test found: The release server runs from a provisioned folder and serves the app, login, page read/write, file listing, and the client bundle over HTTP (all 200), with the page content and created file verified on disk, and all Rust, vitest, and Chromium browser test suites passing (1335, 3023, and 63 tests respectively). 30 minutes.

58SAG-WikiRuns90 / 100Rust1,596

A lightweight knowledge base for people and agents, with ultra fast indexing and most precise retrieval.

What the test found: npm dependencies install, the Vite frontend builds with 0 errors, svelte-check passes with 0 warnings, the sag CLI v0.2.18 binary runs, and 662 vitest tests pass across 86 test files; 8 test files fail due to jsdom ESM incompatibility on Node 18, and 3 files have runtime failures (crypto undefined in worker threads... 42 minutes.

59selfgraphRuns90 / 100Python221

Build and query versioned knowledge graphs for AI agents, with any graph database.

What the test found: SelfGraph 0.2.2 installs cleanly on Python 3.12 via venv and runs its full test suite on embedded Kuzu with all deterministic tests passing. The CLI, SDK quickstart, and MCP server all function correctly. One test is OOM-bound in this container (test_kuzu_cancel_then_close). 48 minutes.

60poznoteRuns86 / 100JavaScript1,050

Powerful note-taking without the hassle.

What the test found: Poznote v6.72.0 PHP application runs on static PHP 8.3 via built-in server; SQLite master and user databases initialized; full REST API v1 test suite (81 tests) passes; login page serves correctly. 36 minutes.

61arscontextaRuns75 / 100Shell3,515

Claude Code plugin that generates individualized knowledge systems from conversation. You describe how you think and work, have a conversation and get a...

What the test found: Ars Contexta v0.8.0 Claude Code plugin builds from source with all 15 kernel primitives validated, all 7 shell scripts syntax-correct, all 39 config files valid, all 5 hooks tested working (one line fix applied to session-orient.sh for zero-session edge case), 249 methodology claims, 3 presets, 17 feature blocks, 16... 15 minutes.

62onedevRuns73.3 / 100Java15,372

The Unified and Autonomous Development Platform

What the test found: OneDev builds and compiles from source with all 86 tests passing, and the server starts on localhost:6610 returning HTTP 200 with the setup initialization page. 38 minutes.

63gollumRuns60 / 100Ruby14,350

A simple, Git-powered wiki with a local frontend and support for many kinds of markup and content.

What the test found: Gollum 6.1.0 builds and runs on MRI Ruby 3.3.5 with rugged/gollum-lib: the server answers HTTP 200 on the root page and renders markdown content, and the full unit test suite passes. 16 minutes.

64llm-wiki-agentRuns60 / 100Python3,727

A personal knowledge base that builds and maintains itself. Drop in sources, Claude (or Codex/Gemini) reads them, extracts knowledge, and maintains a...

What the test found: llm-wiki-agent repo has 9 wiki pages (1 source, 2 entities, 5 concepts, 1 overview), a working knowledge graph (9 nodes, 36 edges), and all 10 Python tools run without errors after fixing 3 bugs in tools/_utils.py and tools/ingest.py. 20 minutes.

65hexo-theme-stellarRuns60 / 100JavaScript2,070

基于 Hexo 的全能个人知识库,开箱即用,内置海量标签和动态数据组件。可用于个人博客、项目文档、知识库、专栏、笔记等。

What the test found: npm install, lint, all 241 unit tests, reuse contract checks, contribution registry checks, and all 4 integration scenarios (post-topic, notebook, wiki, default-content) pass on Node v22.11.0. 7 minutes.

66iweRuns60 / 100Rust1,833

Markdown knowledge graph, LSP for your editor, CLI + MCP memory for your AI agents

What the test found: IWE build 0.24.2 compiles from source, all three binaries (iwe, iwes, iwec) run and respond to commands, and the full test suite passes 606/606 tests. 4 minutes.

67knowledgeRuns60 / 100TypeScript1,464

Knowledge is a tool for saving, searching, accessing, exploring and chatting with all of your favorite websites, documents and files.

What the test found: The repository installs and builds on Linux x86_64 with Node 24.21.0. The bundled Yarn 3.2.4 resolves all workspace dependencies. Angular 14 and Electron 44 build successfully. All 8 unit tests (profile/backup) and all 35 storage service tests pass. The storage service REST API serves /health, /v1/projects CRUD, and... 8 minutes.

68stashbaseRuns60 / 100TypeScript1,378

An open-source, local-first writing workspace where you can write with Claude Code and Codex, drawing on your own sources and past work.

What the test found: StashBase v2.9.7 builds and runs on Ubuntu 24.04 with Node.js 24.21.0 and Python 3.13.16: the Electron app launches real BrowserWindow sessions tested by 6 smoke scripts, the server serves HTTP 200 with the full web UI on port 8090, the Python daemon indexes folders, and all 17 runnable test suites plus the complete... 16 minutes.

69wikmdRuns60 / 100Python423

A file based wiki that uses markdown

What the test found: Wikmd installed in venv, 22/22 tests pass, Flask app serves HTTP 200 with pandoc-rendered wiki content. 14 minutes.

70mulmoclaudeRuns60 / 100TypeScript390

Nurture your own AI assistant on your own computer. Local-first and MIT: memories, data and apps stay as plain files in your workspace. Chat summons the right...

What the test found: Dependencies install, workspace packages build, plugin codegen, 11180 unit tests pass, and the Express API server starts on port 3001 and responds to authenticated API requests. 9 minutes.

71RepoWikiRuns60 / 100Python310

Open-source DeepWiki alternative, generate comprehensive wiki documentation for any codebase from terminal or browser

What the test found: All 134 tests pass, the CLI commands (scan, map, config, serve) execute successfully, the web server starts and serves HTTP 200 on /, and scan produces a complete wiki directory structure even without an LLM API key. 3 minutes.

72ai-trend-studioRuns60 / 100Python240

Local-first AI trend radar and live-session distillation

What the test found: The app installs in a venv, the server answers 200 on /api/health, / and /api/export, refresh collects real AI-relevant trends from all four public sources, and the live-session create/event/finish flow completes with a local summary, with tests 8/8 passing. 6 minutes.

73TextusRuns58 / 100TypeScript814

The next-generation Integrated Writing Environment: documents, AI and plugins side by side. AI keeps your ink flowing; you hold the pen. Free and open source...

What the test found: Under Node v22.23.4 and Rust 1.99.1, the repository's npm install, frontend build, full npm run check, and aicompat Rust tests all pass; the published AppImage launches a Textus window on DISPLAY=:99, and a typed Markdown document renders and serializes correctly in Playwright Chromium. 18 minutes.

74llm-wiki-compilerRuns50 / 100TypeScript2,275

The knowledge compiler. Raw sources in, interlinked wiki out. Inspired by Karpathy's LLM Wiki pattern.

What the test found: npm install, type-check, and build all succeed. 10860 of 10874 tests pass (1300 of 1304 test files). The CLI binary reports version 1.4.0-dev.20260919 and lists all commands. 3 test failures are container-limitation issues: Codex CLI agent live smoke (needs auth), and 2 bwrap isolation tests (needs root to install... 81 minutes.

75mm-wikiRuns with mocks92 / 100Go3,781

MM-Wiki 一个轻量级的企业知识分享与团队协同软件,可用于快速构建企业 Wiki 和团队知识分享平台。部署方便,使用简单,帮助团队构建一个信息共享、文档管理的协作环境。

What the test found: MM-Wiki v0.2.2 builds and runs on Go 1.21.13 with a Python mock MySQL server (port 3308). The application serves HTTP on port 8081, authenticates with admin:123456, and returns the main dashboard at HTTP 200 with the MM-Wiki interface. 34 minutes.

76trustgraphRuns with mocks92 / 100Python3,213

The open-source ontology runtime for AI agents.

What the test found: All 10 TrustGraph Python packages (trustgraph, trustgraph-base, trustgraph-cli, trustgraph-vertexai, trustgraph-flow, trustgraph-bedrock, trustgraph-embeddings-hf, trustgraph-ocr, trustgraph-docling, trustgraph-mcp) installed at version 2.10.0; 3335 unit and contract tests pass; tg-* CLI entry points respond to --help. 11 minutes.

77claude-chat-exporterRuns with mocks92 / 100JavaScript754

One-click, privacy-first exporter for Claude.ai chats, clean Markdown with artifacts and attachments, ready for Obsidian and RAG.

What the test found: The script parses cleanly in Node.js, and all 42 data-pipeline tests pass with mock Claude API response data, covering message ordering via parent chain and index fallback, YAML frontmatter generation, markdown body construction with tables/bold/code, artifact rendering with create/update/reduce, tool_use rendering... 7 minutes.

78OpenTagViewerRuns with mocks92 / 100Java480

Track your AirTags, Apple and Google FindMy devices and custom tags in one place on Android

What the test found: OpenTagViewer APK builds successfully at 107 MB, all 854 JVM unit tests pass, all 680 Python exporter tests pass, all 312 Chaquopy bridge tests pass, and the exporter CLI reports version 1.5.1 and answers --help correctly. 9 minutes.

79wikiRuns with mocks92 / 100Python462

Free and Open Source Wiki built on top of Frappe

What the test found: Python markdown renderer installs and passes all 86 tests; JS SPA frontend, Tailwind CSS theme, highlight.js bundle, and wiki-app.html all build to wiki/public/ and wiki/www/. 18 minutes.

80compound-memoryRuns with mocks92 / 100Python336

Long-term memory for Claude Code and Codex: turns your daily AI sessions into a persistent memory of your preferences, decisions and open threads, injected at...

What the test found: cm doctor, cm snapshot, cm recall, cm distill, cm sessions, and cm --version all return correct output after patching 3 Python bugs and configuring a local mock LLM backend. 8 minutes.

81myPKARuns with mocks92 / 100Python321

AI-powered Personal Knowledge Assistance in a folder. Built on the ICOR methodology. Plain markdown. Any LLM. Yours forever.

What the test found: All 29 Python scripts compile and pass syntax validation. The harness generator (scaffold-init.py) checks and applies successfully, validating 56 generated files against their source hashes. Team validation confirms a compliant structure. All 9 skills pass 146 checks. The expansion-pack test suite passes 21/21 cases... 7 minutes.

82project-cairnRuns with mocks92 / 100Shell255

Turn project work into reusable knowledge, an AI-agent skill for Claude Code & Codex

What the test found: 6 pytest tests pass; Notion graduation pipeline (preflight, single-page graduate, multi-page batch graduate) verified end-to-end via a schema-correct mock on the Notion REST API. 28 minutes.

83terraform-module-releaserRuns with mocks92 / 100TypeScript236

GitHub Action to automate versioning, releases, and documentation for Terraform modules in monorepos.

What the test found: npm install succeeds, TypeScript compiles clean, and 814 of 821 tests pass on Node 24 with terraform-docs v0.21.0 installed to ~/.local/bin/. The 4 skipped/failed suites are GitHub API integration tests requiring a real GITHUB_TOKEN. 4 minutes.

84memoket-kiteRuns with mocks92 / 100Python235

An open-source, vector-free long-term memory engine for AI agents, achieving SOTA on LoCoMo and LongMemEval with significantly less context.

What the test found: memoket-kite 0.1.0 is installed with all dependencies in a Python venv, 425/427 tests pass, and the full Memory remember/recall/answer pipeline executes correctly against a mock OpenAI-compatible provider. 17 minutes.

85kukuRuns with mocks92 / 100TypeScript234

Open-source local-first Markdown workspace with wiki, AI editing, and encrypted sync.

86skill-conductorRuns with mocks92 / 100Python213

Architecture-first skill lifecycle for AI agents. BinEval binary scoring with threshold-blind, cross-family-calibrated judges, gated self-update loop, pressure...

What the test found: All 13 Python scripts in skill-conductor successfully parse and run: smoke tests pass (12/12), structural eval passes (22/22, score 10/10), skill init creates valid structure, packaging produces valid .skill zip, and scripts requiring LLM access (run_eval, improve_description, run_loop) parse their args correctly but... 6 minutes.

87hipocampusRuns with mocks92 / 100JavaScript211

Drop-in memory harness for AI agents, 3-tier memory, compaction tree, hybrid search. One command to set up. Works with Claude Code and OpenClaw.

What the test found: Hipocampus v0.5.3 installs and runs in this container: init creates the full memory directory tree, skills, hooks, config, and platform-specific protocol files; compact processes transcripts through the daily/weekly/monthly compaction chain with secret redaction; uninstall removes all installed artifacts cleanly while... 2 minutes.

88scholar-skillRuns with mocks86 / 100Python227

An OpenClaw skill for academic reading, knowledge linking, reflection, and knowledge evolution in Obsidian.

What the test found: The ScholarSkill setup works: configure.py auto/semi/manual all complete the full flow, artifact verification confirmed byte-identical template sync, correct YAML output, a working placeholder replacement tool, and clean ffmpeg decodes of the packaged images. 2 minutes.

89douyin-to-obsidianRuns with mocks72 / 100Python209

抖音视频与图文知识入库:本地语音转写与文字识别,按主题生成 Obsidian 笔记,维护索引与日志;由 Agent 驱动,CC Connect 为可选接入。

What the test found: Dependencies installed in .venv (103/104 packages, mlx-metal skipped as macOS-only); install script creates working config, vault catalogs, hooks, and slash commands on Linux with platform guard patched; doctor validates all importable modules, paths, and permissions; 121/125 offline tests pass (4 Apple Silicon... 3 minutes.

90nanzi-ai-agent-platformRuns with mocks61 / 100Python247

🚀 开源企业级多智能体平台与 ChatBI 数据洞察平台。支持多 Agent 协作、RAG 知识增强、物理数据源直连与安全沙箱执行、Redis 长期记忆引擎,并集成多模态 EmbedChat 挂件与全方位 Token 消耗审计看板,完整 RBAC权限控制和 Harness支持。

What the test found: 后端依赖已安装;FastAPI 应用已在 SQLite+fakeredis 模拟下启动,/health 返回 200;无基础设施测试中有 231 项通过;由于 Node.js v18.19.1 < 所需版本 v20.19+,前端构建失败;由于容器内不存在 mysql/redis 服务,完整的后端测试套件无法运行. 35 minutes.

91docsRuns with mocks57.3 / 100TypeScript17,003

Docs is an open-source text editor: web-native, made for real-time collaboration, cleanly structured documents and sub-documents with full ownership of your...

What the test found: The Docs backend project installs, builds, and runs 2190+ test cases successfully using mock PostgreSQL, MinIO S3, and compiled email/translation assets, with the only unaddressed failures requiring a real OIDC provider. 62 minutes.

92BookStackRuns with mocks56 / 100PHP19,148

NOW MANAGED ON CODEBERG

What the test found: BookStack PHP dependencies installed, MariaDB 10.11.6 serving on port 3307, all 110+ migrations applied successfully, 1561 of 1622 tests pass or are non-critical failures (LDAP, PDF export, config defaults), the Laravel artisan CLI and development HTTP server both start and respond. 33 minutes.

93beever-atlasRuns with mocks56 / 100Python454

Your First LLM-Wiki Conversation Knowledge Base

What the test found: Beever Atlas v0.3.0 Python package installs and builds fully; all unit tests pass (2 pre-existing env-dependent failures); MCP server starts in stdio mode; test infrastructure works with mocked stores for environments without backing databases. 58 minutes.

94LoreRuns with mocks52.7 / 100TypeScript250

Summon it with a keystroke, throw in anything you want to remember or ask your own memory. A local LLM agent that restructures your knowledge into a vectorized...

What the test found: Install succeeds, unit tests pass (3/3), typecheck and vite build are clean, the Electron app launches a window and runs stable on the virtual display, and a mock-Ollama eval run answers seeded messages end-to-end with all eval endpoints returning HTTP 200. 13 minutes.

95OpenWikiCould not verify40 / 100Rust646

OpenWiki, Mac desktop AI knowledge management tool. Capture clipboard, build personal wiki, get AI insights.

What the test found: OpenWiki 0.3.25 desktop app builds and runs on Linux x86_64 inside the container: the frontend dist compiles, the 489MB debug binary links with WebKitGTK 2.52.6 via GTK-3, and the full application loop starts (clipboard polling, screenshot watching, capture detector, OCR reported as disabled on this platform, wiki... 33 minutes.

96indie-wiki-buddyCould not verify35 / 100JavaScript547

Browser extension that automatically notifies and redirects you to independent wikis.

What the test found: The repository is a browser extension with valid Chrome and Firefox manifests, 846 wiki redirect entries, 33 locale translations, 941 files in the distributable zip, and core JavaScript modules that parse correctly and pass all unit tests. 12 minutes.

-IncarnaMindNot yet tested-TypeScript860-
-obsidian-llm-wikiNot yet tested-TypeScript803-
-swarmvaultNot yet tested-TypeScript738-
-oxynoteNot yet tested-Go680-
-teacher-to-skillsNot yet tested-Python488-
-PersonalWikiNot yet tested-C353-
-obsidian-knowledge-agentNot yet tested-Python203-

runs installed and started with its real dependencies. runs with mocks started after stand-ins replaced external services such as a database or a third-party API. could not verify neither the standard agent nor the stronger one got it running within the time limit; the log shows where it stopped.

How we tested

On this list as of the latest test: 74 projects ran as-is, 20 with mocks, 2 could not be verified, 7 still waiting. Languages tested: Go, Java, JavaScript, PHP, Python, Ruby, Rust, Shell, TypeScript. Every attempt used a clean single-use machine, the subject at a pinned version, and a 90-minute limit for standard projects; the complete procedure is on the methodology page.

Frequently asked questions (FAQs)

How is this list ranked?

By measurement, not opinion: projects Argusic installed and launched on a fresh machine come first, then those that ran with mocks in place of external services, then those it could not verify. Ties go to the Argusic Score, then how popular it is on its own source.

Why are some projects unranked?

7 projects are still waiting for a test or for a finished attempt. They are listed without a rank until Argusic has measured them.

Where is the evidence?

Every row links to the project's Argusic page, where each run has a full log and a terminal recording stored with a sha256 fingerprint. The same pages exist for every one of the tested projects, on this list or not.

More lists in this category

All lists: Best. All tested projects: subjects.

Open source wikis you can self-host, tested | Argusic