{"data":{"slug":"mohitsoni48-turbollm","name":"TurboLLM","tagline":"Run any local LLM engine auto-tuned to your GPU with polished web UI and OpenAI/Anthropic-compatible API","github_url":"https://github.com/mohitsoni48/TurboLLM","owner":"mohitsoni48","repo":"TurboLLM","owner_avatar_url":"https://avatars.githubusercontent.com/u/63787789?v=4","primary_language":"TypeScript","stars":274,"forks":38,"topics":["ai","anthropic-api","claude-code","gguf","gpu","inference","llama-cpp","llama-server","llm","local-llm","offline","openai-api","self-hosted"],"archived":false,"github_pushed_at":"2026-09-19T13:11:45+00:00","maintenance_label":"Very active","stars_delta_30d":49,"url":"https://www.graphcanon.com/tools/mohitsoni48-turbollm","markdown_url":"https://www.graphcanon.com/tools/mohitsoni48-turbollm.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/mohitsoni48-turbollm","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=mohitsoni48-turbollm","description":"Run any local LLM engine, auto-tuned to your GPU — polished web UI + OpenAI/Anthropic-compatible API. Point Claude Code at your own machine in one command. No Electron, no Python, offline-first.","homepage_url":"https://turbollm.dev","license":null,"open_issues":7,"watchers":3,"ai_summary":"TurboLLM provides an efficient and user-friendly way to run local large language model engines optimized for GPU performance, featuring a clean web interface and APIs compatible with major LLM services. It supports self-hosting without dependencies on Electron or Python.","readme_excerpt":"# or install globally\nnpm install -g turbollm\nturbollm\n```\n\n**On first run** the daemon:\n\n1. Detects your GPU and **downloads a matching `llama-server` build** (CUDA for NVIDIA, ROCm\n for AMD, Metal for Apple, SYCL for Intel, Vulkan otherwise — with a CPU fallback).\n2. Starts on and opens your browser.\n3. Walks you through a short **setup wizard**: pick how you'll use it (casual chat, coding,\n tinkering, or pro), get a model recommended for your actual hardware (or bring your own),\n watch it download and load for real, and try it — before landing in **Chat** or **Code**.\n Skippable from any step, and safe to close mid-download — reopen `/onboarding` any time\n from **Models** to pick up where you left off.\n\nAfter that (or any time later), open **Models**, download or pick a GGUF, click **Load**, and\nstart chatting. Stop the daemon any time with **Ctrl+C**.\n\n---\n\n---\n\n## Requirements\n\n- **Node.js 22.13.0 or newer** — enforced at startup with a clear message.\n\n
\n Don't have Node? One command installs it.\n\n
\n\n - **Windows:** `winget install OpenJS.NodeJS.LTS` (or download from )\n - **macOS:** `brew install node` (or download from )\n - **Linux:** use your distro's package manager or — make sure it's\n v22+ (`node --version`)\n\n Then open a **new terminal** (Windows needs one for `PATH` to refresh) and run `npx turbollm`.\n\n
\n\n- **Windows, macOS, or Linux** for `npx turbollm`. A native **Android app** also exists, with a\n bundled Vulkan-accelerated engine — currently in closed/open testing, see turbollm.dev.\n- A GPU is recommended but **not required** — a CPU build is provisioned as a fallback.\n- On Windows, the first time the auto-downloaded `llama-server` runs, SmartScreen/Defender may\n prompt (it's an upstream binary). Allow it once.\n\n---\n\n---\n\n## 🐳 Docker Deployment (NVIDIA / AMD GPU)\n\nTurboLLM includes Docker support with dedicated configurations for NVIDIA CUDA\nand AMD ROCm GPUs.\n\nThe Docker images provide an isolated environment while allowing TurboLLM to\naccess your GPU directly for accelerated local LLM inference.\n\n> The `Dockerfile.*` and `docker-compose-*.yaml` files referenced below live in the\n> [GitHub repository](https://github.com/mohitsoni48/TurboLLM), not in the npm package —\n> clone the repo to use them.\n\n---\n\n## License\n\nSource-available under the **Functional Source License 1.1 (Apache-2.0 future grant)** — SPDX\n**`FSL-1.1-ALv2`**. Full text: [LICENSE.md](https://github.com/mohitsoni48/TurboLLM/blob/main/turbollm/LICENSE.md).\n\n**License FAQ** (plain-language summary — the license text is what's binding):\n\n- **Can I use it for free?** Yes. Personal use, internal business use, education, research,\n and self-hosting — including commercially, inside your company — are all free. The license\n permits *any* use except the one below.\n- **What's the one restriction?** You can't take TurboLLM and ship it (or its functionality)\n as a **competing commercial product or service**. That's the entire boundary.\n- **Can I fork it or redistribute it?** Yes — modify, fork, and redistribute freely for any\n permitted purpose; keep the license text and copyright notices with it.\n- **When does it become fully open source?** The license includes an **irrevocable** grant:\n each release converts to **Apache-2.0 exactly two years** after that release is published.\n The first public release shipped June 2026, so it converts in June 2028 — and every release\n after it follows on its own two-year clock.\n- **Why not MIT/Apache from day one?** TurboLLM is built by a solo maintainer; FSL prevents a\n large vendor from re-skinning it as their own product while it's young, and the future grant\n guarantees the full open-source outcome anyway. (Sentry created the FSL; GitButler, PowerSync,\n and Codecov ship under it too.)\n\n

Built for people who refuse to wa","github_created_at":"2026-06-12T14:41:29+00:00","created_at":"2026-07-15T11:01:17.626514+00:00","updated_at":"2026-09-20T05:07:21.268942+00:00","categories":[{"slug":"inference-serving","name":"Inference & Serving","url":"https://www.graphcanon.com/categories/inference-serving","markdown_url":"https://www.graphcanon.com/categories/inference-serving.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/inference-serving"},{"slug":"model-training","name":"Model Training","url":"https://www.graphcanon.com/categories/model-training","markdown_url":"https://www.graphcanon.com/categories/model-training.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/model-training"}],"tags":[{"slug":"ai","name":"ai"},{"slug":"anthropic-api","name":"anthropic-api"},{"slug":"claude-code","name":"claude-code"},{"slug":"gpu","name":"gpu"},{"slug":"inference","name":"inference"},{"slug":"llama-cpp","name":"llama-cpp"},{"slug":"local-llm","name":"local-llm"},{"slug":"openai-api","name":"openai-api"}],"trust":{"provenance":{"is_fork":false,"github_id":1267439901,"owner_type":"User","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-09-20T05:07:18.750Z","maintenance":{"label":"Very active","score":96,"methodology":"github_public_v1","releases_90d":30,"days_since_push":0,"last_release_at":"2026-09-19T12:55:11Z","stars_delta_30d":49,"open_issues_delta_30d":1},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-15T11:01:18.824Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"mcp":{"source":"repo_scan","observed_at":"2026-09-20T05:07:20.005Z","server_manifest":false},"scan":{"source":"repo_scan","observed_at":"2026-09-20T05:07:20.005Z"},"languages":{"value":["typescript","javascript"],"source":"github.language+package.json","observed_at":"2026-09-20T05:07:20.005Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["When you want to self-host an LLM service without external dependencies on Electron or Python.","For scenarios where offline operation is necessary, ensuring dependency-free access to local LLM engines optimized for GPU."],"when_not_to_use":["If your setup does not include a GPU as TurboLLM primarily optimizes performance specifically for that hardware.","When you require heavy model training capabilities on the same platform; TurboLLM focuses more on running and inference tasks with LLMs."],"source":"enrich:decision_facts","observed_at":"2026-07-17T13:48:59.692Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"TurboLLM offers local LLM execution optimized for GPU performance with a polished web UI and APIs compatible with OpenAI/Anthropic."}]}}