AI Decision System
Replayable, offline-verifiable records for AI-assisted go/no-go decisions: experiment promotions, typed model answers and human-approved actions.
Codex load balancer and proxy for ChatGPT accounts with usage dashboard
Developer toolsPython · TypeScriptGitHub @Soju06OnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
Codex/ChatGPT multiple account load balancer & proxy with usage tracking, dashboard, and OpenCode-compatible endpoints
Taken from the repository README
Loading the full README. The saved excerpt is shown below.
Load balancer for ChatGPT accounts. Pool multiple accounts, track usage, manage API keys, view everything in a dashboard.
Documentation: — getting started, client setup, configuration, deployment, troubleshooting, and more screenshots.
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 upstreamRead the full README ↗English | 简体中文
Load balancer for ChatGPT accounts. Pool multiple accounts, track usage, manage API keys, view everything in a dashboard.
Documentation: https://soju06.github.io/codex-lb/ — getting started, client setup, configuration, deployment, troubleshooting, and more screenshots.
# Docker (recommended)
docker volume create codex-lb-data
docker network inspect codex-lb-net >/dev/null 2>&1 || docker network create codex-lb-net
docker run -d --name codex-lb \
--network codex-lb-net \
-p 2455:2455 -p 1455:1455 \
-v codex-lb-data:/var/lib/codex-lb \
ghcr.io/soju06/codex-lb:latest
# or uvx
uvx codex-lb
# or nix
nix run github:Soju06/codex-lb
Open localhost:2455 → Add account → Done.
Accessing the dashboard remotely for the first time? You need a one-time bootstrap token — see Getting started.
Point any OpenAI-compatible client at codex-lb. For Codex CLI, ~/.codex/config.toml:
model = "gpt-5.6-sol"
model_reasoning_effort = "xhigh"
model_provider = "codex-lb"
[model_providers.codex-lb]
name = "openai" # required — enables remote /responses/compact. Lowercase since Codex 2026-05-23; older "OpenAI" stops resolving gpt-5.5
base_url = "http://127.0.0.1:2455/backend-api/codex"
wire_api = "responses"
supports_websockets = true
supports_standalone_web_search = true # requires codex-lb >= 1.22.0
requires_openai_auth = true # required for codex app
| Logo | Client | Endpoint | Guide |
|---|---|---|---|
| Codex CLI / IDE | http://127.0.0.1:2455/backend-api/codex | Client setup → Codex CLI | |
| OpenCode | http://127.0.0.1:2455/v1 | Client setup → OpenCode | |
| OpenClaw | http://127.0.0.1:2455/v1 | Client setup → OpenClaw | |
| Hermes Agent | http://127.0.0.1:2455/v1 | Client setup → Hermes Agent | |
| OpenAI Python SDK | http://127.0.0.1:2455/v1 | Client setup → Python SDK |
Remote clients need an API key created from the dashboard.
Environment variables with CODEX_LB_ prefix or .env.local — see .env.example and the
configuration guide. SQLite is the default database backend;
PostgreSQL is optional via CODEX_LB_DATABASE_URL.
| Environment | Path |
|---|---|
| Local / uvx | ~/.codex-lb/ |
| Docker | /var/lib/codex-lb/ |
Backup this directory to preserve your data.
Full docs live at https://soju06.github.io/codex-lb/:
Independent projects that consume the dashboard API, maintained outside codex-lb (see the docs listing for access guidance):
# Docker
docker compose watch
# Local
uv sync && cd frontend && bun install && cd ..
uv run codex-lb # backend :2455
cd frontend && bun run dev # frontend :5173
# Nix
nix run .
nix develop # enter dev environment
Thanks goes to these wonderful people (emoji key):
This project follows the all-contributors specification. Contributions of any kind welcome!