BitFan
Public Service Atlas for Bittensor
BitFan
Instant is a developing Bittensor inference service intended to expose OpenAI-compatible streaming requests through miners optimized for serving performance.
The documented architecture places an OpenAI-compatible gateway in front of miners that operate model-serving hardware, while validators measure miner availability and performance and publish Bittensor weights rather than sitting in the customer request path.
The project states that it is still under active development and not yet a production service, and the website currently provides a waitlist rather than public API access.
Developers that will need an OpenAI-compatible low-latency inference endpoint once Instant opens public access.
Instant's documented design separates customer routing at the gateway from miner model serving and validator performance measurement, keeping validators out of the inference request path.
