mnemosyne systems is happy to announce orangu 1.2.0 - the Open Source AI-powered coding environment.
The models worth running locally have gotten a lot larger this year, and running them on your own hardware has mostly meant compromise. As of 1.2.0, orangu handles them. This release also connects orangu to external tools over the Model Context Protocol, and adds two modes that shape its behaviour around what you are actually doing right now.
orangu remains a complete, self-contained stack that runs entirely on your own machine - a terminal environment, an on-demand model manager, and a native GGUF inference server, written end to end in Rust . No llama.cpp, no ggml, no Python. No API keys, no telemetry, and no code leaving your machine.
The big models
The headline of this release is eight new model families: DeepSeek 4, GLM, Kimi K3, Muse-Glimmer, Inkling-Small, Ling, NVIDIA Nemotron, and Qwen 3.8.
Supporting the architectures is only half the job - the machine has to be able to hold them. Two pieces of work in the engine make that practical:
- KV caching, so a long conversation with a large model does not pay the full price again on every turn
- Multi-card support, so a large model can be spread across the GPUs you already have rather than the one you wish you had
Context size suggestions were raised to match, and model names are handled correctly when you refer to a model by number.
MCP support
orangu can now connect to external tools through the Model Context Protocol. Local MCP server management, refresh, and reconnect are all handled inside orangu - no separate supervisor to run, and no editing of a config file to bring a server back after it drops.
Markdown and PlantUML
Markdown rendering in the terminal got a significant upgrade in this release, so answers arrive readable rather than as a wall of source.
In the web console, PlantUML now renders alongside Mermaid. Class diagrams, sequence diagrams, and component sketches come back as diagrams, whichever notation the model reaches for.
Initial license header support
1.2.0 lands a first pass at keeping license headers consistent across a codebase as you work in it. This is the start of the work, not the finished feature - if your project has a header convention orangu gets wrong, this is the release where telling us about it is worth the most.
Developer and Committer mode
Two new modes shape orangu around the task at hand:
/developer- for writing the change/committer- for getting it committed
Same environment, different emphasis, without you having to re-explain what you are up to.
Also in this release
- Model last-use tracking, so it is clear which models on the machine are actually earning their disk space - a community contribution
- Additional answer save formats - a community contribution
orangu-benchin the web console, so a benchmark run no longer means dropping to the command line- An improved
/review - A cheat sheet and endpoint documentation
- macOS and WGPU stability fixes, including recovery from a lost device
- The complete manual ships with the release as PDF and HTML, and is embedded in the binary for offline reading with
/manual
56 commits went into this release, and it again carries community contributions. Our thanks to everyone who filed an issue, opened a discussion, or sent a patch.
Upgrading
The one-liner installer puts the whole stack on the machine in a single step, on Linux, macOS, and Windows:
curl -fsSL https://mnemosyne-systems.github.io/orangu/install.sh | sh
Availability
orangu 1.2.0 is available now for Linux, macOS, and Windows on x86_64 and aarch64, with source and checksums published alongside the binaries. The backends are CPU, Vulkan, Metal, CUDA, ROCm, and OpenCL. Everything is Open Source under the GNU General Public License v3.0 - the environment, the coordinator, and the inference engine alike.
Commercial support for orangu is available from mnemosyne systems - contact sales .
We would genuinely like your feedback - especially from anyone running the larger models, and from anyone on a multi-card setup.
Be part of the community, and be proud of your contributions ! Try it, star it, fork it, break it, tell us what’s missing.
Thanks for your time !
Overview: https://www.mnemosyne-systems.ai/products/orangu
Repository: https://github.com/mnemosyne-systems/orangu
Release: https://github.com/mnemosyne-systems/orangu/releases/tag/1.2.0