Please check the build logs for more information.
See Builds for ideas on how to fix a failed build, or Metadata for how to configure docs.rs builds.
If you believe this is docs.rs' fault, open an issue.
Dotcore
Local AI, without the setup.
Dotcore runs AI on hardware you own. It loads models, fits them to your machines, and puts a single endpoint in front of everything.
- Pick a model, get a running endpoint. Dotcore starts it for you, on llama.cpp, vLLM or ninfer.
- Fit the model to your hardware. See how much memory a model needs and the largest context that fits each machine.
- One endpoint in front of everything. Call every model through one OpenAI-compatible API. If a deployment is down, requests move to another one.
- Many machines, working as one. Start on one machine, then add more.
- Apps next to your models. Run apps on the same machines, for example Open WebUI for chat or ComfyUI for images.
[!IMPORTANT] Dotcore is in alpha. Expect breaking changes, including to saved sites. Security is best effort for now and not guaranteed. See SECURITY.md.
Quick start
It runs on Linux (x86_64, arm64) and macOS (Apple Silicon). You need:
- Nix. It builds Dotcore and every engine the same way on each system, which is what lets one site mix machines and engines.
- Docker Engine on Linux, or Docker Desktop on macOS.
Follow the installer's instructions. See docs/install.md for the full steps.
Documentation
Contributing
Questions, ideas and bug reports go to Discussions. Pull requests fix an accepted issue. See CONTRIBUTING.md.
Security
Report a vulnerability privately, as described in SECURITY.md.
License
Copyright 2026 Dotcore Inc. and contributors. Licensed under Apache-2.0, see LICENSE.
The license covers the code, not the Dotcore name or logo:
- Projects built for Dotcore may say "for Dotcore" or "compatible with Dotcore".
- Modified versions you distribute must use a different name and logo.
- The Dotcore logo is for the official software and website only.