Soju06/codex-lb

Codex-lb is a load balancer and proxy that distributes requests across multiple OpenAI accounts while providing usage tracking and a web dashboard.

The tool addresses the challenge of managing API quotas and rate limits when working with multiple OpenAI accounts. It sits between client applications and OpenAI's API, routing requests across available accounts to balance load and prevent hitting individual account limits. The proxy maintains compatibility with OpenCode-style endpoints, allowing existing integrations to work without modification. Usage tracking is built in, recording consumption metrics across all managed accounts through a dashboard interface.

Teams or individuals managing multiple OpenAI accounts for production workloads should consider this tool if they need transparent quota management and want to avoid service interruptions from hitting per-account limits. It suits scenarios where distributing load across accounts is preferable to requesting higher limits from OpenAI. The project includes OAuth support for authentication and rate-limiting capabilities to control request flow.

The codebase shows active development with regular commits addressing bugs and feature requests. Pull requests receive timely review and merging. The project maintains responsiveness to user-reported issues, with fixes deployed promptly. Documentation is kept current alongside code changes, and the maintainer engages with feature suggestions from the community.