Ollama runs language models locally and provides tools to work with model files and APIs.
Open source
24 alternativesColibri
Colibri runs mixture-of-experts language models on local hardware, streaming experts from disk with a small C engine.
About Colibri
Colibri is an open-source project that runs large mixture-of-experts models on local computers. It keeps the parameters used for each token in RAM and reads expert weights from disk when needed, using C engines that work without a GPU.
Setup checks RAM, disk, CPU and GPU, recommends a model that fits, then builds or downloads the engine and downloads the model with resume support. It requires at least 8 GB of RAM, 22 GB of free disk for the smallest model, and an internet connection, and exposes local OpenAI-compatible and Anthropic-compatible endpoints.
Key features
- Streams MoE expert weights from disk during local inference
- Runs on CPU with optional CUDA or Vulkan GPU support
- Setup script checks hardware and recommends fitting model
- Resumable model downloads with progress display
- Local OpenAI-compatible API at /v1 endpoint
- Local Anthropic-compatible base URL for apps
Best for: People who want to run supported open MoE models on an existing Windows, Linux, or macOS computer.
Good for
- Works without a GPU; CPU operation is supported
- Setup recommends a model based on available RAM and disk
- Model downloads can be stopped and resumed
Watch out for
- Needs at least 8 GB RAM and 22 GB free disk
- Requires an internet connection for setup and downloads
- GPU use depends on operating system, GPU type, and installed packages
Colibri at a glance
- Website
- github.com/JustVugg/colibri
- Type
- Open-source project
- Pricing
- Not confirmed. Check the site
- Open source
- Yes · Source code
- Language
- English
- Last checked
- Oct 5, 2026 · Online when checked
- Listed
- Oct 5, 2026
Colibri alternatives
Tools and sites you could use instead, ranked by how closely they match.
The colibri-core NLP tool and library works with n-grams, skipgrams, and other linguistic patterns.
Open source
24 alternatives
Stakpak runs an agent on a machine to ship code and keep applications running.
Open source
24 alternatives
Glide provides an open model gateway for developing production generative AI applications.
Open source
24 alternatives
Candle provides a minimalist Rust framework for machine learning model development and execution tasks.
Open source
24 alternatives
ONNX-MLIR compiles ONNX models using MLIR compiler infrastructure for model deployment workflows.
Open source
24 alternativesPopular in AI Infrastructure & MLOps
Well-known listings from the same topic.
Guild AIguild.aiFlow AIflow-ai.comsystemprompt.iosystemprompt.ioCoreWeavecoreweave.comKili Technologykili-technology.comAPI docsdocs.x.ai
Colibri FAQ
What hardware is needed to run Colibri?
At least 8 GB of RAM (16 GB or more is better), 22 GB free disk for the smallest model, and an internet connection. A graphics card is optional.
How do I install and start Colibri?
On Windows, download and unzip the repository, then double-click START-HERE.bat. On Linux or macOS, clone the repository and run ./start-here.sh after installing the listed prerequisites, then answer the model question.
Which models are supported?
Ten language-model engines including GLM-5.2/5.3, Kimi K3, DeepSeek V4 Flash, Qwen3.6 and OLMoE, plus Qwen-Image-2.1 for pictures and Laya, GLiNER2.5-Decide and Clef for decisions.
Is something wrong or out of date?
Tell us and an editor will check it.
Own this website?
Claim this listing
Claim the listing and suggest updates yourself. An editor checks every claim first.