Vibe Coding DiscoverMiniSearch
View on GitHubMinimalist web-searching platform with an AI assistant that runs directly from your browser. Demo: https://felladrin-minisearch.hf.space
MiniSearch is a self-hosted private search engine: SearXNG aggregates web results, a local cross-encoder reranks them, and an AI assistant writes cited answers using either in-browser LLMs (WebGPU/CPU via wllama) or an OpenAI-compatible backend. Ships as one Docker container, no API key or telemetry required.
Use Cases
Private web search with no API key or trackingRAG answers with cited sources over search resultsRun small LLMs (135M-4B) locally in the browser on WebGPU/CPUSelf-host a single-container search engineConnect Ollama, LM Studio, vLLM, llama.cpp or AI Horde as backendFollow-up questions with conversation memoryRerank search results locally with a cross-encoder modelSet as the browser default search engineTrigger searches from Raycast quicklinksEvaluate retrieval and answer quality with a golden set
Built With
- Language
- TypeScript
- Frameworks
- React · Vite · Mantine · Vercel AI SDK · wllama · ONNX Runtime · HuggingFace Tokenizers · Dexie · SearXNG · Docker · Vitest · Playwright · Biome
Tags
web-search · rag · in-browser-inference · webgpu · local-llm · privacy · searxng · self-hosted · reranking · citations · embeddings · question-answering · wasm · docker · metasearch · typescript