01 — safe
Link your development tools and start collaborating with AI
02 — flexible
Create requests and break them into manageable tasks
03 — observable
AI agents implement changes with full git tracking
04 — accountable
Maintain complete visibility with automatic documentation
05 — performant
Inspect file history and understand every change
alchemylab
>
press enter to start building
press to start building ↵

Directory Sandboxing

All modifications are restricted to the project directory

Full File History

Complete version history and recoverability for every file change

Branch Isolation

Each request is an isolated branch supporting safe multi-agent workflows

Directory Sandboxing

All modifications are restricted to the project directory

Full File History

Complete version history and recoverability for every file change

Branch Isolation

Each request is an isolated branch supporting safe multi-agent workflows

1. Create Request
request
2. Break Down
task
task
task
3. Make Changes
edit
4. Git Commit
commit

Client Portability

Take your workspace from client to client without losing context

MCP Integration

Plugs in as a Model Context Protocol server for seamless AI tooling

Right Tool for the Right Job

Turn on the tools you want to use. And off the ones you don't

Client Portability

Take your workspace from client to client without losing context

MCP Integration

Plugs in as a Model Context Protocol server for seamless AI tooling

Right Tool for the Right Job

Turn on the tools you want to use. And off the ones you don't

A
A
A
A
A
A
Codex CLI
MCP Client
Cursor
Claude Code
opencode
observable
Agent Activity Timeline
See exactly what your agent has been doing
Created file
2:34 PM
src/auth/login.ts
Modified function
2:35 PM
validateUser()
Ran tests
2:36 PM
18 passing
Updated docs
2:37 PM
README.md
Committed changes
2:38 PM
5 files changed
Task completed
2:39 PM
Authentication system

Real-time Activity

Watch your AI agents work in real-time with live updates. See every file creation, modification, test run, and commit as it happens. The timeline shows you exactly what your agent is doing at any moment, giving you complete visibility into the development process.

Track progress, understand decisions, and maintain full awareness of your agent's actions from start to finish.

accountable

Complete Code Accountability

Track every line of code back to its originating request. View detailed histories, inspect request metadata, and maintain full transparency across your entire codebase. Try it in the editor!

src/api/user-service.ts
Attribution
History
Audit
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
import { User } from './types';
import { db } from './database';
export class UserService {
async authenticate(email: string, password: string) {
const user = await db.findUser(email);
return user && user.verifyPassword(password);
}
async getProfile(userId: string): Promise<User> {
const user = await db.findById(userId);
if (!user) {
throw new Error('User not found');
}
return user;
}
async updateProfile(userId: string, data: Partial<User>) {
const updated = await db.update(userId, data);
return updated;
}
async deleteUser(userId: string) {
await db.delete(userId);
}
}

Request Timeline

req-user-auth-system

2:47 PM
Request completed
Merged into main branch
2:45 PM
Modified validateUser()
src/auth/validator.ts · +12 -3 lines
2:40 PM
Created authentication middleware
src/middleware/auth.ts · +45 lines
2:34 PM
Created JWT utilities
src/utils/jwt.ts · +67 lines
2:30 PM
Request created
Add JWT authentication to user service

Parallel Execution

Native multi-agent support enables concurrent task execution on any number of agents

Building API
Writing Tests
Documentation
Code Review

Serial Excellence

Built from the ground up with advanced context management in mind. Finetuned tooling that builds on evaluations of real software development

35K
Context Window
-68%
Tool Calls
2.3s
Response Time