Debian Launches Free AI Inference Portal for Its Developers

The Debian project has launched a new service, the Debian Inference Portal, giving contributors self-service access to LLM inferencing. The initiative follows the project's recent vote to permit the "responsible use of generative AI."

At launch, two models are available through the portal: DeepSeek V4 Flash and GLM-5.2. The service exposes an API that developers can wire into their preferred code editors and other AI tooling. Obtaining an access key requires a Debian Salsa account, so the portal is open to Debian Developers and Debian Maintainers.

The project is funded by Scaleway, which sponsors a fixed monthly pool of credits, with a weekly budget allocated per user. That arrangement keeps costs bounded while spreading the resource across contributors.

The inference is meant for practical, project-oriented work: packaging, bug triage, fixes, tooling, documentation, and related tasks. The service itself is hosted at inference.debian.net.

The move is notable because Debian — one of the oldest Linux distributions, with a reputation for caution and conservatism — has not merely allowed its developers to use AI but has gone a step further and handed them a ready-made, paid-for infrastructure to do so.

Source: Phoronix