Pumpkin: open-source GUI manager gets llama.cpp running in 5 minutes, no terminal needed
gbolgurov · reddit · 2026-10-27
A developer open-sourced Pumpkin, an all-in-one graphical manager for llama.cpp built for his girlfriend who can't use a terminal: it detects your hardware, downloads the right llama.cpp build, and picks a quant that fits your GPU/RAM — install to running model in about 5 minutes (screenshot shows Qwen3.6 35B Q6K at 41.7 tok/s, 14.8/16GB VRAM).
Feature highlights:
- Auto-discovers existing models from LM Studio and Ollama
- Chat with web search, tools, MCP servers, Claude-style skills, and connectors for GitHub, Notion, Git, Postgres and more
- Pumpkin Code: a coding agent that works in a folder and can ask before changing files or running commands
- Q&A over your own documents; a small vision model describes images for models that can't see
- Image/short-video generation via stable-diffusion.cpp (optional install), usable from your phone
Early release, best tested on Ubuntu and Mac; Windows and Android work but are rougher. Free and open source, non-commercial.
More from coding & agent
- DevSpeak MCP lets your coding agent polish PR descriptions and commit messages — modelcontextprotocol · 2026-10-27
- ThreadMind MCP gives AI agents a Git-like memory tree to cut token costs — mahmu_tech · 2026-10-27
- He let Claude design a complete e-ink business card, from schematic to product page — isaul32 · 2026-10-27
- Papers with Code revives as an Apache 2.0 open-source project with public API and MCP server — NielsRogge · 2026-10-27
- The AI engineer job in 2023 vs 2026: one viral blank line — JafarNajafov · 2026-10-27
- DeepSeek Flash impresses dev: let it build its own tools, cheap enough to experiment — victormustar · 2026-10-27