Introduction

The Inference.net OSS Partner Program helps open-source maintainers capture the value they create through their applications. By integrating Inference.net as an LLM provider option in your BYOK (Bring Your Own Key) application, you can earn a share of the profits generated when users choose Inference.net as their LLM provider.

Many open-source applications rely on third-party LLM inference providers like OpenAI, Anthropic, OpenRouter, and Inference.net. Typically, users configure these applications with their own API keys. While this “Bring Your Own Key” approach allows developers to build products without charging users directly, it has become difficult for open-source maintainers to monetize their work—most value accrues to the API providers who get paid for inference requests.

The Inference.net OSS Partner Program changes this dynamic by sharing profits with the developers who drive these inference requests, creating a sustainable stream of income for open-source project maintainers.

How It Works

  1. Integrate: Add Inference.net as an LLM provider option in your application
  2. Configure Keys: Your users select Inference.net and configure their API keys
  3. Revenue Generation: When users make inference requests through your app, we track usage
  4. Profit Sharing: You receive 50% of profits after costs (approximately 5-10% of total revenue)
  5. Payouts: Withdraw your earnings once you reach the $100 threshold

Benefits for Partners

  • Passive Income Stream: Get paid for the inference requests your users make through your application
  • Transparent Dashboard: Track usage and revenue metrics in real-time
  • Marketing Support: Get promotional assistance from Inference.net to your users
  • No Exclusivity: Continue supporting multiple LLM providers alongside Inference.net
  • No Long-Term Commitments: Join or leave the program at your discretion

Code Integration

Inference.net offers full OpenAI API compatibility, making integration straightforward:

  1. Add Inference.net as a provider option in your application settings
  2. Configure the base URL to point to Inference.net’s API endpoints
  3. Add the required header to inference requests using the OpenAI SDK
  4. Provide documentation for your users on obtaining and configuring Inference.net API keys

Example integration code:

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.inference.net/v1",
    api_key=os.environ.get("INFERENCE_API_KEY"),
    default_headers={
        "X-Partner": "<YOUR_PARTNER_ID>"
    },
)

response = client.chat.completions.create(
    model="meta-llama/llama-3.1-8b-instruct/fp-8",
    messages=[{"role": "user", "content": "What is the meaning of life?"}],
    stream=True,
)

for chunk in response:
    if chunk.choices[0].delta.content is not None:
        print(chunk.choices[0].delta.content, end='', flush=True)

Available Models

Check our models catalog for a complete and up-to-date list of models. We regularly add new models as they are released. Our most popular models are:

Becoming a Partner

To join the Inference.net OSS Partner Program:

  1. Send an email to [email protected] with your project’s GitHub link, a brief description, and information about current usage.
  2. Submit required information:
    • Link to your project on GitHub
    • Brief description of your project
    • Information about current usage
  3. Our team will review your application and reach out with approval
  4. Once approved, implement the integration and start earning

Frequently Asked Questions

Eligibility & Application

Who can apply to the Partner Program?
Any developer of an open-source application that supports BYOK (Bring Your Own Key) functionality can apply. There are no specific licensing requirements or minimum usage thresholds, though your application must be able to integrate with OpenAI-compatible APIs.

What’s the application process like?
Send an email to [email protected] with your project’s GitHub link, a brief description, and information about current usage. Our team will review your application and contact you regarding approval.

Revenue & Payments

How exactly is the profit-sharing calculated?
Partners receive 50% of profits after costs. Since Inference.net pays approximately 80% of revenue to GPU providers, partners typically receive between 5-10% of the total revenue generated through their applications.

How and when are payments distributed?
Payments are currently made weekly via ACH once you’ve accumulated at least $100 in earnings. Inference.net is working to automate this process, which will eventually allow partners to withdraw funds every 24 hours or set up automated payments directly to their bank accounts.

Are there any tax implications I should be aware of?
Payments are made as donations where applicable. Partners are responsible for reporting and paying appropriate taxes according to their local regulations. Inference.net does not provide tax advice, so we recommend consulting with a tax professional.

Technical Integration & Support

What technical changes do I need to make to integrate with Inference.net?
Integration requires adding Inference.net as a provider option in your application, setting the correct base URL, and adding a partner header to inference requests made through the OpenAI SDK. We provide example code and documentation to simplify this process.

Does Inference.net provide technical support for integration issues?
While partners are generally responsible for building and maintaining their integrations, our team can provide technical guidance and support where possible. We offer documentation and example implementations to facilitate smooth integration.

How complete is the OpenAI API compatibility?
Inference.net offers full OpenAI API compatibility, including advanced features like function calling, JSON mode, and other OpenAI-specific functionalities. This makes integration straightforward for applications already supporting OpenAI.

User Experience & Tracking

How does this affect my users?
Users will need to create an Inference.net account to obtain an API key, which they’ll then configure within your application. The actual inference process remains unchanged from their perspective, though they may benefit from Inference.net’s competitive pricing.

How can I track usage and revenue?
Partners have access to a dashboard showing usage and revenue metrics. While general information about generations is publicly visible, only partners can see the exact amounts accrued and withdraw funds.

Is there a minimum revenue threshold for withdrawals?
Yes, partners need to accumulate at least $100 in earnings before they can withdraw funds.

Business Terms

Is exclusivity required?
No, there are no exclusivity requirements. You can continue to support other LLM providers alongside Inference.net, giving your users freedom of choice.

Are there any long-term commitments?
No long-term commitments are required. Partners can join or leave the program at their discretion.

How does Inference.net’s pricing compare to other providers?
Inference.net is positioned as one of the most affordable and performant open-source model providers on the market. If your users are already using open-source models, they will likely save money by switching to Inference.net.

What kind of marketing support does Inference.net provide?
Inference.net helps partners promote their integrations to our existing user base and through social media channels. This can help drive adoption of your application and increase your revenue through the partner program.

Performance & Reliability

What can I expect in terms of uptime and reliability?
Inference.net models run on a distributed network of data centers, providing better uptime and reliability than many traditional providers. This architecture ensures consistent performance for your users.


For additional questions or to apply to the Partner Program, contact our partner support team at [email protected].