{"id":50139873,"url":"https://github.com/andersondanieln/hexllama","last_synced_at":"2026-05-24T00:33:58.154Z","repository":{"id":358777405,"uuid":"1240150652","full_name":"andersondanieln/hexllama","owner":"andersondanieln","description":"A beautifully crafted desktop client for running and managing local LLMs via llama.cpp.","archived":false,"fork":false,"pushed_at":"2026-05-23T16:55:15.000Z","size":1678,"stargazers_count":14,"open_issues_count":2,"forks_count":3,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-05-24T00:33:52.956Z","etag":null,"topics":["ai-assistant","gguf","llama","llama-cpp","llm","local-ai","machine-learning"],"latest_commit_sha":null,"homepage":"","language":"TypeScript","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/andersondanieln.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-05-15T20:25:56.000Z","updated_at":"2026-05-23T16:55:18.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/andersondanieln/hexllama","commit_stats":null,"previous_names":["andersondanieln/hexllama"],"tags_count":7,"template":false,"template_full_name":null,"purl":"pkg:github/andersondanieln/hexllama","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/andersondanieln%2Fhexllama","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/andersondanieln%2Fhexllama/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/andersondanieln%2Fhexllama/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/andersondanieln%2Fhexllama/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/andersondanieln","download_url":"https://codeload.github.com/andersondanieln/hexllama/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/andersondanieln%2Fhexllama/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":33417487,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-23T22:14:44.296Z","status":"ssl_error","status_checked_at":"2026-05-23T22:14:43.778Z","response_time":53,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai-assistant","gguf","llama","llama-cpp","llm","local-ai","machine-learning"],"created_at":"2026-05-24T00:33:57.611Z","updated_at":"2026-05-24T00:33:58.141Z","avatar_url":"https://github.com/andersondanieln.png","language":"TypeScript","funding_links":[],"categories":[],"sub_categories":[],"readme":"\u003cdiv align=\"center\"\u003e\n  \u003cimg src=\"assets/github-logo-hexllama.png\" alt=\"Hexllama Logo\" width=\"400\" /\u003e\n\u003c/div\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003cimg src=\"https://img.shields.io/github/v/release/andersondanieln/hexllama?style=flat-square\u0026color=black\u0026label=version\" alt=\"Latest Version\" /\u003e\n  \u003cimg src=\"https://img.shields.io/badge/Electron-191970?style=flat-square\u0026logo=Electron\u0026logoColor=white\" alt=\"Electron\" /\u003e\n  \u003cimg src=\"https://img.shields.io/badge/React-20232A?style=flat-square\u0026logo=react\u0026logoColor=61DAFB\" alt=\"React\" /\u003e\n  \u003cimg src=\"https://img.shields.io/badge/TypeScript-007ACC?style=flat-square\u0026logo=typescript\u0026logoColor=white\" alt=\"TypeScript\" /\u003e\n  \u003cimg src=\"https://img.shields.io/badge/Vite-B73BFE?style=flat-square\u0026logo=vite\u0026logoColor=FFD62E\" alt=\"Vite\" /\u003e\n\u003c/p\u003e\n\n\u003cbr/\u003e\n\nHexllama is a fast, native desktop interface designed to streamline managing and running local Large Language Models using llama.cpp. It strips away the friction of command-line execution and manual file management, providing a unified workspace to discover, download, configure, and serve models.\n\nBuilt by and for local AI enthusiasts, Hexllama ensures you spend less time wrestling with terminal arguments and more time interacting with models.\n\n## Features\n\n**Integrated Model Hub**\nSearch Hugging Face directly within the application. Browse repositories, view file details, and download GGUF models with a single click without ever opening a browser.\n\n![Model Hub](assets/screenshots/model-hub.png)\n\n**Smart Download Manager**\nPause, resume, or cancel large model downloads reliably. You can also paste direct GGUF links. When a download completes, Hexllama automatically generates an execution template with recommended threads, batch sizes, and context windows tailored to the model's parameters and quantization level.\n\n![Model Download](assets/screenshots/model-download.png)\n\n**Template-Based Execution**\nSave your configurations as reusable templates. Run multiple models simultaneously on different ports without conflict. Launch them in \"Chat UI\" mode to automatically open the built-in llama.cpp web interface, or \"API Only\" mode to serve them silently in the background.\n\n![My Templates](assets/screenshots/my-templates.png)\n\n![Template Settings](assets/screenshots/template-edit-model-settings-parameters.png)\n\n**Version and Backend Management**\nRunning cutting-edge models sometimes requires different builds of llama.cpp. Hexllama lets you maintain and seamlessly switch between multiple backend binaries. It automatically checks the ggml-org repository for new releases and lets you download and extract them straight from the settings panel.\n\n**Visual Command Editor**\nStop memorizing execution flags. Edit backend-specific commands through a structured user interface. Toggle booleans, set limits on numerical inputs, and define default parameter values for the llama.cpp server.\n\n![Settings](assets/screenshots/settings.png)\n\n## Installation\n\n### Download the Release\nThe fastest way to get started is to use the pre-compiled installer.\n\n1. Navigate to the [Releases](https://github.com/andersondanieln/hexllama/releases) page.\n2. Download the installer for your operating system.\n3. Run the installer and launch Hexllama.\n\n### Run Locally\nIf you want to build from source or modify the project, you can easily run the development environment.\n\nPrerequisites:\n- Node.js 18 or higher\n- npm\n\n```bash\n# Clone the repository\ngit clone https://github.com/andersondanieln/hexllama.git\n\n# Enter the project directory\ncd hexllama\n\n# Install dependencies\nnpm install\n\n# Start the development server\nnpm run dev\n```\n\nTo compile the application into an executable for your current OS:\n```bash\nnpm run build\n```\n\n## Roadmap\n\n### Phase 1: Core Foundation (Completed)\n- [x] **Integrated Model Hub**: Hugging Face search \u0026 download direct from the app.\n- [x] **Smart Download Manager**: Pause/resume/cancel, auto-template generation based on hardware \u0026 quant level.\n- [x] **Template-Based Execution**: Run multiple models on different ports, reusable configuration templates.\n- [x] **Version and Backend Management**: Download and switch between different versions of `llama.cpp` binaries directly.\n- [x] **Visual Command Editor**: Graphical UI for configuring server parameters instead of terminal flags.\n\n### Phase 2: Enhanced Inference \u0026 Native UI (Short to Mid-Term)\n- [ ] **Built-in Chat Interface**: Native chat client to interact with models directly within Hexllama without launching external browser tabs.\n- [ ] **MTP (Multi-Token Prediction) Support**: Enable faster generation speeds using speculative decoding / MTP.\n- [ ] **TurboQuant Support**: Support optimized quantizations and execution configurations.\n- [ ] **Multi-Language Support**: Complete internationalization (i18n) to support Portuguese, English, Spanish, etc.\n\n### Phase 3: Multi-Backend \u0026 Advanced Engines (Long-Term)\n- [ ] **Alternative Backend Integration**: Expand support beyond `llama.cpp` to include:\n  - **MLX**: Native backend for Apple Silicon optimized performance.\n  - **vLLM / ExLlamaV2**: Support for high-throughput and GPU-optimized engines.\n\n## Acknowledgements\n\nThis project exists because of the incredible foundational work of Georgi Gerganov and the ggml-org community. Please consider supporting the development of [llama.cpp](https://github.com/ggerganov/llama.cpp).\n\n## Privacy and Terms\n\nHexllama is provided as is, without warranty of any kind. The developers assume no liability for damages or issues arising from the use of this software.\n\nThis application is strictly local. It does not collect, store, or transmit any telemetry or personal data. Note that downloading models relies on third-party services like Hugging Face, and executing backends relies on the downloaded binaries, both of which are subject to their own respective privacy policies.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fandersondanieln%2Fhexllama","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fandersondanieln%2Fhexllama","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fandersondanieln%2Fhexllama/lists"}