Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/21st-dev/1code/llms.txt

Use this file to discover all available pages before exploring further.

1Code brings together multiple AI coding agents in one unified interface. Switch between Claude Code, Codex, and custom models without changing tools or workflows.

​
Why It Matters

Different agents have different strengths:
  • Claude Code excels at complex refactoring and architectural changes
  • Codex offers advanced planning with structured proposals
  • Custom models let you bring your own API keys and providers
With multi-agent support, you can:
  • Use the right agent for each task
  • Compare approaches from different models
  • Switch providers without losing context
  • Bring your own API keys (BYOK) for any provider

​
Available Agents

​
Claude Code

Claude Code is Anthropic’s official coding agent, integrated directly into 1Code. Setup:
  1. Go to Settings → Accounts
  2. Click “Connect Claude Code”
  3. Sign in with your Anthropic account or import an existing token
Features:
  • Full access to Claude 3.5 Sonnet and other Anthropic models
  • Native tool support for file editing, bash commands, and web search
  • OAuth integration for secure authentication
  • Multi-account support for switching between workspaces
Implementation: The Claude Code router (src/main/lib/trpc/routers/claude-code.ts:109) manages OAuth flows and token storage using Electron’s secure storage.

​
Codex

Codex is an AI coding agent from Zed Industries, featuring advanced planning capabilities. Setup:
  1. Go to Settings → Accounts
  2. Click “Connect Codex”
  3. Sign in with your ChatGPT account or use an API key
Features:
  • GPT-5.3 Codex model with high-quality reasoning
  • Built-in plan mode with structured proposals
  • API key authentication for direct access
  • Session persistence across restarts
Implementation: The Codex router (src/main/lib/trpc/routers/codex.ts:1) uses the Agent Communication Protocol (ACP) to communicate with the Codex binary.

​
Custom Models (BYOK)

Bring your own API keys to use any OpenAI-compatible provider. Supported Providers:
  • OpenAI (GPT-4, GPT-4o, o1)
  • Anthropic (Claude via API key)
  • Google Gemini
  • Any OpenAI-compatible endpoint
Setup:
  1. Go to Settings → Models
  2. Add your API key
  3. Configure base URL (for custom providers)
  4. Select your model

​
Switching Between Agents

​
In the UI

1

Open Model Selector

Click the model name in the chat header or press Cmd+K
2

Choose Your Agent

Select from Claude Code, Codex, or custom models in the dropdown
3

Continue Chatting

Your conversation continues with the new agent. Previous messages remain in context.

​
Per-Chat Configuration

Each chat can use a different agent. This is useful for:
  • Testing the same prompt with multiple models
  • Delegating specific tasks to specialized agents
  • Comparing output quality across providers

​
Agent Marketplace

Custom agents can be loaded from:
  • User agents: ~/.claude/agents/
  • Project agents: .claude/agents/ in your repo
  • Plugin agents: Installed via the plugin marketplace
Example Agent File: 
---
name: code-reviewer
description: Reviews code for best practices and security issues
model: claude-3-5-sonnet-20241022
tools:
  - read
  - grep
  - bash
---

You are an expert code reviewer. Focus on:
- Security vulnerabilities
- Performance optimizations
- Code maintainability
- Best practices for the given language
Agents are loaded dynamically from the filesystem (src/main/lib/trpc/routers/agents.ts:17) and can be referenced with @agent-name in chat.

​
Best Practices

Choose the Right Agent

  • Use Claude Code for general coding tasks and refactoring
  • Use Codex when you need structured planning before execution
  • Use custom models for specialized domains or cost optimization

Compare Outputs

Fork a chat at any message to try the same prompt with a different agent.

Set Project Defaults

Configure default agents per project in .claude/config.json