loci
Local-first harness and ticket operations for AI assistants via CLI or MCP.
README
<p align="center"> <img src="packages/web/src/assets/logo.svg" alt="Loci Logo" width="128" /> </p>
🗂️ Loci
Local-first harness and ticket operations. Run it in any project, track work with your AI assistant via CLI or MCP.
Why Loci?
Loci is a lightweight local-first harness system that runs entirely on your machine.
Your AI coding assistant (Claude, Gemini, Cursor, etc.) can read and update workflow tickets directly via loci CLI or MCP.
Install
Recommended one-line install (macOS Apple Silicon + Linux x86_64/aarch64):
curl -fsSL https://raw.githubusercontent.com/thienhm/loci/main/scripts/install.sh | bash
Install a specific version:
curl -fsSL https://raw.githubusercontent.com/thienhm/loci/main/scripts/install.sh | bash -s -- v2.0.1
Transition bootstrap (legacy Bun launcher):
bun install -g github:thienhm/loci
loci update
During the transition, loci update refreshes both the managed Rust binary and the global Bun package that still provides loci serve and loci open.
Managed binary location: ~/.loci/bin/loci
What's New In 2.0.1
Loci 2.0.1 is a patch release for the Rust-primary 2.0 line:
loci updatenow refreshes both the managed Rust binary and the global Bun wrapper used byloci serveandloci open.- Dashboard reads now recover from local SQLite open edge cases and keep healthy registered projects visible.
- Ticket creation no longer reports a hard failure when only the post-create registry summary refresh fails.
- CLI help and release documentation now describe the
loci update/loci upgradeboundary more clearly.
Quick Start
# In your project directory
loci init
# Optional: start the server + web UI
loci serve
# Optional: open the web UI in your browser
loci open
The web UI is available at http://localhost:3333 by default.
Connect Your AI Assistant
CLI (Recommended)
Use the local CLI directly in your agent workflow. loci init generates project instructions and foundation docs:
AGENTS.mdLOCI.mdloci/project.mdloci/architecture.mdloci/validation.mdloci/guardrails.mdloci/current-state.mdloci/glossary.mdloci/backlog.md
MCP
Loci also exposes an MCP server at http://localhost:3333/mcp (requires loci serve to be running).
Claude Desktop / Claude Code
{
"mcpServers": {
"loci": {
"url": "http://localhost:3333/mcp"
}
}
}
Gemini CLI (~/.gemini/settings.json)
{
"mcpServers": {
"loci": {
"url": "http://localhost:3333/mcp"
}
}
}
Cursor / Windsurf
{
"loci": {
"url": "http://localhost:3333/mcp"
}
}
CLI Reference
Project
| Command | Description |
|---|---|
loci init |
Initialize Loci in the current project |
loci doctor |
Check harness/project health |
loci serve |
Start the MCP server and web UI |
loci open |
Open the web UI in your browser |
loci update |
Update the managed Rust binary and transitional Bun wrapper |
loci upgrade |
Upgrade project templates/data |
Tickets
| Command | Description |
|---|---|
loci list [--json] |
List workflow tickets |
loci add "title" [--json] |
Create a workflow ticket |
loci get <id> [--json] |
Get one ticket by ID |
loci status <id> <status> [--json] |
Update ticket status |
loci patch <id> [--assignee] [--progress] [--priority] [--labels] [--json] |
Update ticket fields |
loci plan, loci ready, loci validate, loci evidence, loci trace, loci summary, loci review |
Harness workflow commands |
Docs & Attachments
| Command | Description |
|---|---|
loci doc read <id> <filename> [--json] |
Read a ticket document |
loci doc write <id> <filename> --content "..." |
Write a ticket document |
loci attachments <id> [--json] |
List attachments for a ticket |
How It Works
- Project docs live under
loci/(human/agent-visible context and workflow packets) - Operational state lives under
.loci/(SQLite + local tool state) - The CLI is local-first and does not require a running server for core workflows
- The MCP server and web UI are served by
loci serve
Updating
loci update
This installs the managed Rust binary and refreshes the global Bun loci package so TypeScript fallback commands such as loci serve use the latest server/web code. If the wrapper refresh fails, the command prints the manual recovery command:
bun remove -g loci && bun install -g github:thienhm/loci
If you're on an older install and loci update cannot run yet, bootstrap once manually:
bun remove -g loci && bun install -g github:thienhm/loci
Then run:
loci update
Migration Notes
loci synchas been retired. Useloci upgrade.loci skillhas been retired from product CLI.
Contributing
See CONTRIBUTING.md.
License
MIT — see LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.