llampart 1.0.0 brings polished local web UI to llama-server with multi-language and MCP support
New open-source UI enhances local LLM chats with translations, MCP tools, and a Frosted Glass theme.
llampart 1.0.0 is a new standalone local web UI designed specifically for llama-server, the inference engine from the popular llama.cpp project. Built by developer mossy_troll_84, it evolved from the original llama-ui into a separate, more polished interface focused on everyday local use. The UI offers a complete desktop-style experience with extended settings panels for appearance, model configuration, MCP (Model Context Protocol) integration, tools, and data management. It supports six languages out of the box: English, Polish, German, French, Italian, and Spanish. The two-column conversation sidebar includes advanced features like date/time display, conversation pinning, selective deletion, and a delete-all function that preserves pinned chats. Users can also import and export conversations locally while avoiding sensitive settings leakage.
The interface offers three theme modes: dark, light, and a distinctive Frosted Glass style. Bundled wallpapers and wallpaper customization let users personalize their workspace. For advanced users, llampart includes MCP-related UI flows for managing servers, tools, resources, and prompts, as well as a minimal Reasoning/Tools display mode. Deployment flexibility comes with an optional Caddy guide for local or LAN setups. The project is MIT-licensed and built with Svelte/SvelteKit, with careful attribution to llama.cpp. This is an initial public release, so the developer welcomes feedback and issue reports. For professionals running local LLMs, llampart bridges the gap between bare-bones terminal interfaces and heavy hosted solutions, offering a clean, feature-rich environment for extended chats.
- Multi-language interface supporting English, Polish, German, French, Italian, and Spanish for broader accessibility.
- Advanced conversation management with pinning, selective deletion, and delete-all while preserving pinned chats.
- Extended settings include MCP integration, tools, data sections, and three visual themes (dark, light, Frosted Glass).
Why It Matters
Local LLM users get a polished, feature-rich UI that streamlines daily workflows without relying on cloud services.