For Open Source

Free inference for
open source builders.

Building open source with AI? We'll give you up to 2M tokens/month for free. Powered by solar energy California wastes.

OSS Builder
500Ktokens/month
500,000 tokens on Priority tier, completely free
Requirements
  • Public GitHub repo with OSI-approved license
  • Active commits in last 90 days
  • Non-commercial project
What you get
  • 500K tokens/month on Priority tier
  • All models (Llama 70B, 8B, DeepSeek-V3)
  • Streaming support
  • Optional "Powered by Daylite" badge
Apply Now
OSS Maintainer
Most popular
2Mtokens/month
2,000,000 tokens on Priority tier, completely free
Requirements
  • 5,000+ GitHub stars on the project
  • OSI-approved license (MIT, Apache, BSD, etc.)
  • Active maintenance (commits in last 60 days)
  • Non-commercial use
What you get
  • 2M tokens/month on Priority tier
  • All models including future releases
  • Priority support via email
  • Early access to new features
  • Green Inference certificate for your project
Apply Now
How to Apply
1.

Email oss@daylite.ai with a link to your GitHub repo

2.

We verify the license, activity, and star count (takes 1-2 business days)

3.

You get an API key with your OSS tier credits — no credit card, no term limits

4.

Optionally add a "Powered by Daylite" badge to your README (not required, but appreciated)

The Badge (Optional)

If you want to show your project runs on green AI inference, add this to your README:

[![Powered by Daylite](https://daylite.ai/badge.svg)](https://daylite.ai)

FAQ

Can I use this for a commercial project?

The OSS program is for non-commercial open source projects. If your project has commercial users, check our regular pricing — it's already competitively priced with per-customer cost tracking included.

Is there a time limit?

No term limits. As long as your project stays active and open source, your credits renew monthly.

Do I have to add the badge?

No. The badge is optional and appreciated, not required. We believe in earning your endorsement, not demanding it.

What models can I use?

All models: Llama 3.1 70B, Llama 3.1 8B, DeepSeek-V3, and all future additions.

Build open source. We'll power the inference.

Apply for the OSS Program