Skip to main content
We work with inference providers and model developers that can offer something valuable to NanoGPT users. If you think you would be a good fit, contact us at support@nano-gpt.com or through our Discord.

​
What we look for

We receive many integration requests, so we prioritize providers that give users a clear reason to choose them. That could include:
  • Models that are not widely available through other providers
  • A meaningful advantage in price, latency, throughput, reliability, privacy, geographic coverage, or infrastructure
  • A distinctive technical approach, such as specialized hardware or a decentralized backend
  • Sufficient production capacity and operational support to handle sustained traffic reliably
Size alone is not a requirement. A smaller provider can still be a strong fit if it offers a distinctive model or capability. However, if the offering is otherwise similar to providers we already support, we generally expect meaningful operational scale. A single-GPU or single-operator service is unlikely to be a fit unless it offers something genuinely unique.

​
Stealth model launches

We are also open to launching models in stealth, including under a temporary or non-public name. A stealth launch can help providers and model developers learn how people use a model in practice before a broader release—for example, for programming, creative writing, role play, or general conversation. If you would like to test a model this way, please reach out. We can discuss the scope, duration, presentation, and technical requirements of the launch.

​
What to include when contacting us

To help us evaluate your offering, please include:
  • Your company name
  • The best way to contact you—we generally prefer email
  • Why you think we should integrate your service
    • Do you offer models that are difficult to access elsewhere?
    • Do you have especially competitive pricing or volume discounts?
    • Do you offer a meaningful technical, privacy, performance, or infrastructure advantage?
  • Your available infrastructure, approximate capacity, regions, and operational support
  • Your rate limits
  • Your privacy policy, including whether prompts are logged and how long they are retained
  • Whether you support credit card payments or automatic invoicing
  • Whether you provide an OpenAI-compatible endpoint
  • Whether you are proposing a public integration or a stealth model launch
  • Any other information that may be relevant