Multi-Account ChatGPT OAuth for OpenCode

Use multiple ChatGPT Plus/Pro personal or organization accounts with OpenCode. Never hit rate limits again.

┌────────────────────────────────────────────────────────────────┐
│                                                                │
│   Account 1 (rate limited) ──┐                                 │
│   Account 2 (rate limited) ──┼──► Auto-rotate ──► Keep coding  │
│   Account 3 (available) ─────┘                                 │
│                                                                │
└────────────────────────────────────────────────────────────────┘

Why Multi-Account?

| Problem | Solution | |---------|----------| | Hit ChatGPT rate limits while coding | Add multiple accounts, auto-rotate when limited | | Team members share one subscription | Each person adds their own account | | Different orgs have separate subscriptions | Use accounts from multiple organizations | | One account gets throttled | Seamlessly switch to next available account |

Quick Start

# Install
npx -y opencode-openai-multi-auth@latest

# Add your first account
opencode auth login
# Select "ChatGPT Plus/Pro (Codex Subscription)"

# Add more accounts (optional but recommended)
opencode auth login
# Select "Add Another OpenAI Account"

# Start coding - accounts rotate automatically on rate limits
opencode run "write hello world to test.txt" --model=openai/gpt-5.2 --variant=medium

How Multi-Account Works

Adding Accounts

# First account
opencode auth login
# → Select "ChatGPT Plus/Pro (Codex Subscription)"
# → Browser opens, login with ChatGPT
# → Account saved

# Second account (different email/org)
opencode auth login
# → Select "Add Another OpenAI Account"
# → Login with different ChatGPT account
# → Account added to rotation pool

# Repeat for as many accounts as you have

Automatic Rotation

When you hit a rate limit:

1. Plugin detects 429 (rate limited) response
2. Marks current account as limited for that model
3. Switches to next available account
4. Retries your request automatically
5. Shows toast notification: Switched to [email protected]

Account Selection Strategies

| Strategy | Behavior | Best For | |----------|----------|----------| | sticky (default) | Stay with one account until rate limited | Single user, predictable usage | | round-robin | Rotate through accounts on each request | Distribute load evenly | | hybrid | Sticky within session, rotate across sessions | Multiple terminal sessions |

Set via environment variable:

OPENCODE_OPENAI_STRATEGY=round-robin opencode run "task"

Team Usage

Each team member can add their own ChatGPT account:

# Developer 1 adds their account
opencode auth login  # logs in as [email protected]

# Developer 2 adds their account  
opencode auth login  # → "Add Another OpenAI Account" → [email protected]

# Developer 3 adds their account
opencode auth login  # → "Add Another OpenAI Account" → [email protected]

All accounts are pooled - when one person's account is rate limited, the plugin uses the next available.

Environment Variables

| Variable | Description | Default | |----------|-------------|---------| | OPENCODE_OPENAI_QUIET=1 | Disable toast notifications | Off | | OPENCODE_OPENAI_DEBUG=1 | Enable debug logging | Off | | OPENCODE_OPENAI_STRATEGY | Account selection strategy | sticky | | OPENCODE_OPENAI_PID_OFFSET=1 | Offset account selection by PID | Off |

Account Management

View Accounts

cat ~/.config/opencode/openai-accounts.json | jq '.accounts[] | {email, planType}'

Remove All Accounts

rm ~/.config/opencode/openai-accounts.json

Check Rate Limit Status

cat ~/.config/opencode/openai-accounts.json | jq '.accounts[] | {email, rateLimitResets}'

Slash Commands (TUI)

/codex-status

Shows usage status for all configured accounts.

Models

All GPT-5.2 and GPT-5.1 models with reasoning variants:

• gpt-5.2 (none/low/medium/high/xhigh)
• gpt-5.2-codex (low/medium/high/xhigh)
• gpt-5.1-codex-max (low/medium/high/xhigh)
• gpt-5.1-codex (low/medium/high)
• gpt-5.1-codex-mini (medium/high)
• gpt-5.1 (none/low/medium/high)

Note: The model selector reflects what the ChatGPT OAuth backend advertises. API-only models (like gpt-5-mini/nano) may not appear until the backend exposes them.

Configuration

• Modern (OpenCode v1.0.210+): config/opencode-modern.json
• Legacy (v1.0.209 and below): config/opencode-legacy.json

# Modern install
npx -y opencode-openai-multi-auth@latest

# Legacy install
npx -y opencode-openai-multi-auth@latest --legacy

# Uninstall
npx -y opencode-openai-multi-auth@latest --uninstall

Features

• Multi-account rotation - Add unlimited ChatGPT accounts, auto-rotate on rate limits
• Per-model rate tracking - Each model's limits tracked separately per account
• Toast notifications - Visual feedback when accounts switch
• OAuth authentication - Same secure flow as official Codex CLI
• 22 model presets - All GPT-5.2/5.1 variants pre-configured
• Automatic token refresh - Never manually re-authenticate
• Multimodal support - Image input enabled for all models

Documentation

• Getting Started
• Configuration Guide
• Troubleshooting
• Architecture

Credits

Maintained by ZenysTX Original implementation by Numman Ali Inspired by opencode-google-antigravity-auth

Usage Notice

This plugin is for personal development use with your own ChatGPT Plus/Pro subscriptions.