codex-lb¶
Load balancer for ChatGPT accounts. Pool multiple accounts, track usage, manage API keys, view everything in a dashboard.
Features¶
- Account pooling — load balance across multiple ChatGPT accounts
- Usage tracking — per-account tokens, cost, 28-day trends
- API keys — per-key rate limits by token, cost, window, model
- Dashboard auth — password + optional TOTP
- OpenAI-compatible — Codex CLI, OpenCode, any OpenAI client
- Auto model sync — available models fetched from upstream
Where to go¶
- Getting Started — Docker / uvx quick start, remote bootstrap token
- Client Setup — Codex CLI, OpenCode, OpenClaw, Python SDK
- Configuration — the few settings that matter
- Anonymous Telemetry — collected fields, consent, disabling, and retention
- Authentication — dashboard auth modes
- Conversations — dashboard view and conversation APIs
- API Keys — protecting proxy routes
- Routing — routing strategy guide
- Database — SQLite / PostgreSQL, data paths, Postgres upgrades
- Deployment — Docker, Kubernetes, remote access
- Troubleshooting
Screenshots¶
| Settings | Login |
|---|---|
| Dashboard (dark) | Accounts (dark) | Settings (dark) |
|---|---|---|
codex-lb is spec-driven: normative behavior lives in OpenSpec capabilities in the repository. Docs pages describe how to use the project and link back to the specs that govern them.