# Lampson — open-source AI coding agent for your terminal and browser

> Markdown rendition of https://lampson.org/en — the same page a browser renders as HTML.

A coding agent for your projects

# Point it at a folder.
It reads, edits and runs — *and asks first.*

Lampson is a coding agent you install once and open inside any project. Talk to it in the terminal or in your browser, watch every step it takes, and say yes or no before anything risky happens. It works with the model you already pay for.

[Install](https://lampson.org/en/install)
[Read the docs](https://lampson.org/en/docs)
[GitHub](https://github.com/kitecosmic/lampson)

$`npm i -g lampson`

then `cd` into a project and type `lampson`

my-app — lampson

```
lampson 0.2 · deepseek-chat · build · ask

the login form doesn't show errors when the password is wrong. fix it

✓ grep "password" src/                    3 files
✓ read src/components/LoginForm.tsx       88 lines
✓ read src/api/auth.ts                   41 lines
✓ edit src/components/LoginForm.tsx       +6 −1
   40   } catch (err) {
   41 +   setError("Wrong email or password");
? bash npm test                          allow? ❯ yes  no
✓ bash npm test                          14 passed

The form now shows “Wrong email or password” under the
password field when the API returns 401. Tests pass.

─── 5 steps · 12k tokens · 38s
```

## Nothing to learn. Nothing to configure.

You describe what you want in plain words. Lampson figures out which files matter, changes them, runs the tests, and shows you the result — one visible step at a time.

### 01Stays inside your project

Lampson can only touch the folder you opened it in. Not your home directory, not the project next door. This isn't a setting you could forget — it's how the language it's written in works.

### 02Asks before anything risky

Reading and editing just happen. Deleting things, installing packages, touching git history, running installers — those stop and wait for your *yes*. Truly destructive commands are refused outright.

### 03Terminal or browser

`lampson` opens a chat in your terminal. `lampson --web` opens the same agent in your browser, with a file tree, live logs of the servers it started, and a real terminal tab.

### 04Your model, your key

DeepSeek, GLM, Kimi, Groq, OpenRouter, Ollama on your own machine, Anthropic, OpenAI… Pick a provider on first run, paste a key, done. Switch any time with one command.

### 05Remembers your project

How to run it, what broke last time, decisions you made together — Lampson keeps notes per project and reads them back next session. You can read and edit those notes too.

### 06Works while you're away

“Every morning at 9, run the tests and tell me.” Scheduled tasks run on their own; if one hits something risky, you get a link on your phone to approve or deny it.

Plugins

## Give it new abilities. Turn them on when you want.

A **plugin** is a small folder that teaches Lampson a new tool: query your database, call your company's API, deploy, send a message. Plugins are **off by default** — you switch each one on, and only then can the agent use it.

- A plugin is a folder with a `plugin.json` and a script in any language.

- A Synsema plugin runs under a capability ceiling; an executable one runs as is — either way, turning it on is your call.

- The agent can *write* a plugin for you — but never turn it on by itself.

- Keep plugins global (every project) or inside one project.

[How plugins work →](https://lampson.org/en/docs/plugins)

plugins

```
$ ls ~/lampson/plugins/
example-hello/  deploy/  postgres/

$ cat ~/lampson/plugins/postgres/plugin.json
{
  "name": "postgres",
  "description": "read-only queries on the dev database",
  "kind": "exec",
  "command": "python query.py",
  "tools": [{ "name": "query", "readonly": true, … }]
}

lampson ❯ /plugins on postgres
☼ postgres is on — 1 tool available: plugin_postgres_query
```

Also in the box

## The things you'd expect, without the setup.

- **Sub-agents** — it can split a job in parallel workers and gather the results.

- **Skills** — drop a `SKILL.md` in your project and the agent learns your procedures.

- **MCP servers** — the same JSON config you already use elsewhere.

- **Language servers** — go-to-definition and references instead of grep guessing.

- **Sessions and traces** — every step is saved; reopen a conversation days later.

- **Images** — paste a screenshot and ask about it.

Works with

DeepSeekAnthropicOpenAIGLM (Z.ai)KimiGroqGrokOpenRouterMiniMaxOllamaany OpenAI-compatible API

Your API key is stored on your machine and only ever sent to the provider you chose. Lampson has no server of its own — nothing leaves your computer except your conversation with the model.

## Try it on a project you have open right now.

One line to install. One word to start.

$`npm i -g lampson`

[Other ways to install](https://lampson.org/en/install)
[Quickstart](https://lampson.org/en/docs/quickstart)