Pneuma: An AI-Generated Desktop Environment Where Software Materializes from Descriptions
Pneuma is a desktop computing environment where software doesn't exist before you need it. You boot to a blank screen with a prompt, describe what you want—like a CPU monitor, game, notes app, or data visualizer—and a working program materializes in seconds.
How It Works
Your input goes to an LLM, which generates a self-contained Rust module. That gets compiled to WebAssembly in under a second, then JIT-compiled and executed in a sandboxed Wasmtime instance. Everything is GPU-rendered via wgpu (supporting Vulkan, Metal, and DX12). If compilation fails, the error is automatically fed back for correction, with a ~90% first-attempt success rate.
Architecture and Features
- Microkernel architecture: Agents run in isolated WASM sandboxes with a typed ABI for drawing, input, storage, and networking
- Agent persistence: Once generated, agents persist—you can reuse them, they communicate through IPC, and you can share them through a community agent store
- System stability: An agent crash can't bring down the system
- Agent capabilities: Agents can run side by side, persist to a local store, and be shared or downloaded from the community store
Current Status and Future Goals
Currently runs as a desktop app on Linux, macOS, and Windows. The longer-term goal is to run on bare metal and support existing ARM64 binaries alongside generated agents, creating a full computing environment where AI-generated software and traditional applications coexist. Built entirely in Rust.
Philosophy
The creator built this because they believe the traditional software model—find an app, install it, learn it, configure it—creates unnecessary friction. If a computer can generate exactly the tool you need when you need it, and keep it around when it's useful, why maintain a library of pre-built software at all?
A free tier is available (no credit card required), and there's a video on the landing page showing it in action. The creator is interested in feedback on the concept, UX, and whether this is something developers would actually use.
📖 Read the full source: HN LLM Tools
👀 See Also
Qwen 3.6 27B Quantization Benchmark: Q4_K_M Beats Q8_0 on Practical Tradeoffs
Evaluated Qwen 3.6 27B across BF16, Q4_K_M, and Q8_0 GGUF quants on HumanEval, HellaSwag, and BFCL. Q4_K_M delivers near-BF16 scores with 48% less RAM, 1.45x speed, and 68.8% smaller file size.
MoltSoup: A Persistent Multiplayer World for AI Agents to Compete
MoltSoup is a persistent multiplayer environment where AI agents can explore six zones, fight monsters, trade via an order-book market, and engage in PVP. Agents interact by reading a skill.md file and making HTTP calls to the API.
Pair Programmer Plugin Adds Live Screen, Voice, and Audio Context to Claude Code
A developer has built a plugin called Pair Programmer that gives Claude Code real-time desktop perception by capturing screen, microphone, and system audio streams. The architecture uses specialized agents running in parallel for different input types, with indexing currently handled by cloud models but designed to be model-agnostic.
SideX: A Tauri-Based Port of Visual Studio Code
SideX is a port of Visual Studio Code that replaces Electron with Tauri, using a Rust backend and the OS's native webview. The project claims the same architecture with 96% smaller size, with core editing and terminal functionality currently working.