kylrix
OfficialTL;DR
Kylrix — open-source, local-first workspace for notes, goals, workspaces, chat, vault, and agents.
Use the cloud — kylrix.space
Self-host —
curl -fsSL https://raw.githubusercontent.com/Kylrix/kylrix/master/selfhost.sh | bash→ app on:5003Wire an agent — mint a PAT (your workspace) or agent key (the agent's own workspace), then
npx skills add kylrix/kylrix --skill mcp --skill api --skill agents
Related MCP server: Shared Workspace MCP
Sponsor
Self-host
Agent skill:
npx skills add kylrix/kylrix --skill selfhostConfigure (optional — omit to auto-mint admin credentials into .env):
export SELFHOST_ADMIN_EMAIL=you@example.com
export SELFHOST_ADMIN_PASSWORD='your-secure-password'Install (bundled Appwrite + Kylrix, no cloud backend):
curl -fsSL https://raw.githubusercontent.com/Kylrix/kylrix/master/selfhost.sh | bashDefault | |
App |
|
Appwrite API |
|
More: SELFHOST.md · re-run ./selfhost.sh anytime (detects drift, skips healthy steps).
Develop
git clone https://github.com/Kylrix/kylrix.git
cd kylrix
cp env.sample .envInstall Ota (execution contract for this repo):
curl -fsSL https://dist.ota.run/install.sh | shRun:
ota doctor
ota up --workflow dev # local app → http://localhost:3005
ota up --workflow verify # lint + test + buildContract: ota.yaml · schema: appwrite.config.json
Humans & agents
Humans and agents share the same workspace. MCP for IDE tool loops; REST for scripts, mobile, and CI.
Choose auth
Token | Use when |
PAT ( | The agent acts in your workspace (IDE tools, scripts, MCP on your behalf). Settings → Developers |
Agent key ( | The agent gets its own workspace — it provisions itself and mints its own PAT. Settings → Smart Agents |
Surface | Use when |
WebMCP ( | In-browser agents (Chrome, ChatGPT browser) with zero-config live session tools |
MCP | IDE agents (Cursor, Claude, Windsurf, Codex, …) |
REST API ( | Scripts, mobile apps, CI, custom backends |
Steps
Mint a token — PAT (your workspace) or agent key (agent workspace)
Install skills
npx skills add kylrix/kylrix --skill mcp --skill api --skill agentsConnect MCP (IDE only — uses your PAT; Smithery wires the official endpoint)
npx -y @smithery/cli install kylrix/kylrix --client cursorBrowser Agents (WebMCP) — Auto-exposed via
navigator.modelContextwith zero config in your live session:await navigator.modelContext.executeTool('kylrix_create_note', { title: 'Agent Note', content: 'Created via in-browser modelContext', tags: ['webmcp'] });Test: Enable
chrome://flags/#enable-webmcp-testingin Chrome, browse via ChatGPT, or click the WebMCP badge in Settings → Developers.
Wiring reference: docs/integrations.md · docs/webmcp.md
Integrations
Link | |
WebMCP (W3C in-browser) | docs/webmcp.md · |
MCP | Humans & agents above · docs/mcp.md |
REST API | docs/api.md · |
Sign in with Kylrix (OAuth 2.1) | |
SDK |
|
Flows | Extensible layers inside Kylrix — kylrix.space/flows |
What ships in the box
Area | What |
Notes & ideas | Linked notes, tags, sharing |
Goals | Goal tracking and focus sessions |
Events | Calendar and scheduling |
Forms | Build forms and collect responses |
Flows | Installable workflow plugins (kylrix.space/flows) |
Workspaces | Projects, collaborators, permissions |
Connect | Chats, moments, group hangouts |
Vault | Client-encrypted credentials (optional) |
Agents | In-workspace sessions with tool parity to users |
Local copy is the default source of truth; sync confirms in the background.
Feedback & security
This server cannot be deployed
Maintenance
Related MCP Connectors
Coding agents build full-stack apps in persistent workspaces and share them by link.
Deploy and manage your apps, databases, storage, and scheduled jobs from your AI agent
- DartOAuthcom.dartai
AI-native project management for tasks, docs, collaboration, and agents.
Project management for AI agents: tasks, docs, decisions and time in one shared team context.
Related MCP Servers
- AlicenseAqualityDmaintenanceFull AI agent lifecycle platform — create, grade, optimize, deploy and orchestrate AI agents from your IDE or terminal.316 npm1MIT
- AlicenseNot gradedqualityCmaintenanceLocal-first memory, pipelines, learning, feedback, and safe code tools for AI coding agents.MIT
- AlicenseBqualityDmaintenanceEnables non-coders to build software by describing ideas in natural language, orchestrating AI agents and CLI workers via chat without needing a terminal.307 npm4Apache 2.0
- AlicenseCqualityBmaintenanceEnables AI coding agents to navigate massive codebases through fast code property graph queries, sandboxed recursive language model execution, durable semantic memory, and swarm concurrency coordination.4MIT