1Code

1Code is an open source visual client for coding agents that combines Claude Code and Codex in one app. It supports desktop and web workflows, Git-based review and merge actions, and background agent runs in cloud sandboxes.

AI Code Assistant AI Developer Tools Programming Agent

Visit Website

Overview

1Code is an open source visual client for coding agents. It brings Claude Code and Codex into one app so developers can choose the agent that fits a task, then manage the work from a calm interface instead of the terminal.

The product is built around long, parallel coding sessions. It supports desktop and web access, Git-based review and merge flows, worktree isolation for separate sessions, and background agents that keep running in isolated cloud sandboxes.

Core features

Multiple coding agents

Switch between Claude Code and Codex from the same app so different tasks can use different agent strengths.

Parallel agents

Run several sessions at once, each in its own context, for parallel feature work.

Built-in Git workflow

Stage files, inspect diffs, commit changes, and open pull requests without leaving the app.

Worktree isolation

Keep each session isolated in its own git worktree to reduce conflicts and avoid stashing.

Background agents and previews

Continue work in isolated cloud sandboxes even when your laptop is closed, with live browser previews for branches.

Integrations and triggers

Connect MCP servers and trigger agents from GitHub, Linear, or Slack using @1code mentions and automations.

Common use cases

Choose the right agent for the task
Use one agent for deep codebase understanding and another for fast execution, without switching products.
Parallel feature development
Run separate sessions for unrelated features so work can continue in parallel with isolated context and worktrees.
In-app code review and PR creation
Review diffs, stage changes, and open PRs from the same interface when you want a visual replacement for terminal-heavy Git workflows.
Background work that outlives your laptop
Keep an agent working overnight in a cloud sandbox, then resume locally with the branch and context intact.
Agent automation from team tools
Trigger automations from GitHub, Linear, or Slack to handle PR reviews, task follow-up, or issue fixes.

Pros and Cons

Pros

Supports both Claude Code and Codex in one interface.
Combines agent execution with visual Git, diff, and pull request management.
Offers parallel sessions and worktree isolation for cleaner multitask workflows.
Includes background agents with cloud sandboxes and live browser previews.
Can connect to MCP servers and trigger work from GitHub, Linear, or Slack.

Cons

The public pages do not spell out detailed platform or system requirements beyond desktop support and a web/PWA option.
Pricing and plan details are partially unclear because the dedicated pricing URL returns a 404, although the site does show Open Source, Pro, and Max tiers on the homepage.
The source leaves several implementation details open, including the exact authentication flow and which MCP servers or external tools are already preconfigured.

FAQ

What is 1Code?

1Code is a visual client for coding agents. It lets you connect an agent account, run work from the desktop app or web, and manage edits, diffs, and pull requests in one interface.

Which coding agents are supported?

The source specifically names Claude Code by Anthropic and Codex by OpenAI as supported agents. It also says you can bring your own API key, and the FAQ asks about supported agents more broadly.

Is 1Code open source?

Yes. The site describes 1Code as open source and includes a self-host it yourself option in its pricing section.

How does authentication work?

You can sign in with Anthropic or OpenAI, or bring your own API key. The site does not provide more detailed authentication mechanics beyond that.

How is 1Code different from the CLI?

1Code is presented as a visual client with Git integration, worktrees, and PR workflows, while the site also has a separate Automations page for triggers and PR review automation. The homepage positions it as an app for running and managing agents, not just a command-line interface.