Beyond Ollama
A Windows desktop AI application built on a heavily customised open source model engine. We added file access, tools, MCP connections, guardrails and hooks, so the model can do work on your machine instead of only talking about it.
- Windows 11
- x64
- Go
- llama.cpp / CMake
- MCP
Under active development. Follow the build log for progress.
Our builds are finished and gated, but they are not yet signed with a publicly-trusted certificate. Windows would present them as an unknown publisher. We are not shipping software that teaches you to click past the warning that protects you. Releases open the day signing is in place.
Local models run beautifully in the apps built to run them, then become a nightmare the moment you want to connect them to anything else. We kept hitting that wall, so we turned the problem around.
- They already run well inside the engine, so we decided to work inside the engine.
- It was open source, so we took it and modified it.
- We gave it file access, tools and MCP connections.
- Then guardrails and hooks, to keep all of that inside limits you set.
It is not a frontier coding agent, but it is far closer than it has any right to be, and it gets better every week.
Local models that can actually touch your files.
In active build.
We list these because you will find them anyway. A limitation you learn about after paying is a defect in how we sold it to you.
Documentation
Installation, requirements, configuration and licensing for Beyond Ollama.
Release history
Every release, what changed in it, and the hash of exactly what shipped.