I Run My Own Inference Server: Can I Plug It Into This Aggregator and Earn USDC?

Can You Plug Your Own Inference Server Into an Aggregator?

If you run your own inference server for models like Claude, GPT, DeepSeek, Qwen, GLM, or Kimi, you might wonder whether you can contribute that capacity to an aggregator and earn USDC. The short answer is: it depends on the aggregator's design, but many aggregators support contributor models where you supply an API key or endpoint. This article explains the general concept and what the earning flow typically looks like.

How Contribution Typically Works

Aggregators that accept contributors usually provide a way to register your inference endpoint or API key. Once approved, the aggregator can route requests to your server. You remain responsible for the underlying infrastructure, model weights, and uptime. The aggregator handles user authentication, billing, and request distribution.

Key points:

• You provide an API key or a reachable endpoint (e.g., an HTTP API) that conforms to the aggregator's expected interface.
• The aggregator may perform health checks, latency tests, or validate model responses before sending production traffic.
• You set your own capacity limits, such as maximum concurrent requests or rate limits, to avoid overloading your server.
• The aggregator tracks usage per request and attributes it to your contributor account.

The Earning Flow

When a user makes a request through the aggregator and it is routed to your server, the aggregator records the usage. Contributors are typically credited based on the official price of the model, with a multiplier. For example, a common model is:

• Contributors earn official price × 1.1 (or × 1.2 for premium models) in USDC.
• Users pay official price × 1.3.

The difference covers the aggregator's operational costs and margin. Payouts are usually made in USDC, often on a periodic basis (e.g., weekly or monthly) once a minimum threshold is reached. Some aggregators may require you to set a payout wallet address.

What You Need to Get Started

Before you can contribute, ensure you have:

• A stable, publicly reachable inference endpoint (or an API key from a supported provider).
• Documentation of your endpoint's API (request/response format, authentication method).
• Monitoring and logging to track usage and troubleshoot issues.
• A USDC wallet address on a supported chain (e.g., Base) to receive payouts.
• Agreement to the aggregator's terms of service, including any uptime or quality requirements.

Practical Considerations

Contributing your inference server is not entirely hands-off. Consider these factors:

• Uptime and reliability: If your server goes down, the aggregator may route requests elsewhere, and you may lose earnings or be penalized.
• Cost management: You pay for your own compute, bandwidth, and any underlying API costs. Ensure the contributor rate (official price × 1.1 or × 1.2) covers your expenses and provides a margin.
• Security: Expose only what is necessary. Use authentication, rate limiting, and network controls to protect your server.
• Compliance: You are responsible for complying with the model provider's terms of service and any applicable laws.
• Scalability: If demand grows, you may need to scale your infrastructure to handle increased load.

Is It Worth It?

Whether contributing makes sense depends on your costs, spare capacity, and the aggregator's contributor rates. If you already run inference servers and have idle capacity, contributing can turn that into USDC earnings. However, if your costs exceed the contributor rate, you may lose money. Always do the math based on your own infrastructure expenses.

Next Steps

If you want to explore contributing, start by checking whether the aggregator has a contributor program. Review the technical requirements, apply, and test with a small amount of traffic before scaling up. Keep an eye on your usage and earnings dashboard to ensure the arrangement is profitable for you.