Vercel
RecommendedThe quickest path: Vercel Functions with durable Upstash Redis and QStash scheduling.
- Free*
- 1-click deploy
- 2 min setup
Tibo giveth, we sure must taketh.
Put every Codex reset to use.
Your always-on service automatically redeems eligible full resets before they expire – even while you or your computer sleeps.
Deploy almost anywhere
Vercel
Docker
Railway
Render
Fly.io
Hetzner
AWS
Google Cloud
Azure
DigitalOcean
Oracle Cloud
Koyeb
Coolify
Portainer
Dokploy
Vercel
Docker
Railway
Render
Fly.io
Hetzner
AWS
Google Cloud
Azure
DigitalOcean
Oracle Cloud
Koyeb
Coolify
Portainer
Dokploy
Vercel
Docker
Railway
Render
Fly.io
Hetzner
AWS
Google Cloud
Azure
DigitalOcean
Oracle Cloud
Koyeb
Coolify
Portainer
Dokploy
Vercel
Docker
Railway
Render
Fly.io
Hetzner
AWS
Google Cloud
Azure
DigitalOcean
Oracle Cloud
Koyeb
Coolify
Portainer
Dokploy
Docker
Run it on your own always-on server with one Bun container and one persistent volume.
- Your server
- One container
- One volume
Codex Auto Reset uses private APIs
Use at your own risk. This independent project uses undocumented OpenAI endpoints and an unofficial device-login flow. OpenAI can change or disable them at any time. Your deployment is user-owned and OAuth credentials are encrypted at rest.
Open source and self-hosted
The useful details.
Codex Auto Reset is a focused automation service for Codex full-reset credits. It runs in infrastructure you control and exists for one narrow job: redeem an eligible reset before it expires, even when your own computer is offline.
What exactly does Codex Auto Reset redeem?
It monitors the full-reset credits reported for your ChatGPT account and redeems an eligible credit close to its expiry. It does not spend credits arbitrarily or change your normal usage window unless OpenAI accepts the redemption. The selected credit and a stable redemption ID are persisted before the request, reducing the risk of duplicate consumption after an ambiguous response.
Where are my OpenAI credentials stored?
Your deployment performs device authorization directly with OpenAI. OAuth credentials are encrypted at rest with your own master key and remain inside infrastructure you control – Upstash Redis for the Vercel edition or the persistent volume for Docker. Codex Auto Reset has no shared backend, central account database, or credential proxy.
Should I choose Vercel or Docker?
Vercel is the quickest setup and combines serverless functions with Upstash Redis and QStash scheduling. Docker runs as one Bun process in one container with one persistent volume, making it suitable for Railway, Render, Fly.io, a VPS, or a home server. Connect only one active redeemer to an OpenAI account at a time.
What happens when OpenAI or the network is unavailable?
Scheduled checks retry throughout the configured redemption window. Rotated OAuth credentials are saved before follow-up requests, and an existing redemption ID is reused after an ambiguous failure. The service keeps its durable state between invocations or restarts so a temporary provider error does not silently create a second redemption request.
Is this an official OpenAI product?
No. Codex Auto Reset is an independent open-source Beta project. It relies on private, undocumented APIs and an unofficial device-login flow that OpenAI may change or disable without notice. Enable device code authorization for Codex in ChatGPT Security settings before setup, and use the project at your own risk.
Support this project
Help maintain this project and its creator. Your support is greatly appreciated!
Coming soon: