Ross ROSS = Recommend OSS · open-source software intelligence for agents

dtsola/xiaoyaosearch

小遥搜索,听懂你的话、看懂你的图,用AI找到本地任何文件。让搜索像聊天一样简单。XiaoyaoSearch: Understands your words, reads your images, finds any local file with AI. Making search as easy as chatting. observed · 2026-08-28

github.com/dtsola/xiaoyaosearch · homepage · Python · NOASSERTION (other) observed · 2026-08-28

Health v2 · maintenance only

70/100

  • Activity 83
  • Release rhythm 81
  • Longevity 21

Flags: no_license

How is this computed?

round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-02. Adoption (stars, forks) is never an input.

  • gap_med: 6.5
  • age_days: 295
  • days_rel: 131
  • days_push: 103
  • n_releases_24m: 13

Full methodology

Adoption not part of the score

1083 stars · 60 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

XiaoyaoSearch is a cross-platform desktop application (Electron + Python/FastAPI) that lets users find local files using AI-powered semantic search with text, voice, or image queries. It combines hybrid retrieval (Faiss vector search + Whoosh full-text) with models like BGE-M3, FasterWhisper, and CN-CLIP to index and search documents, audio, and video entirely on the local machine, and exposes itself as an MCP server and Agent Skill for AI assistants.

Use cases

  • find any file on my computer by describing it in natural language
  • search the contents of pdf, office, and markdown documents locally with AI
  • search local files by speaking instead of typing
  • locate a video or audio file by what is said or shown in it
  • find files using an image as the search query
  • let Claude Desktop or Cursor search my local files through MCP
  • run private semantic file search offline without uploading my data
  • improve recall for domain jargon and synonyms when searching documents

When to choose

  • you need AI-powered search over your own local files with strong privacy (local-first, optional cloud APIs)
  • you want multimodal querying - voice, text, or image - against documents, audio, and video
  • you work with Chinese-language content and want models like BGE-M3/CN-CLIP plus a terminology library
  • you want to give AI assistants (Claude Desktop, Claude Code, Cursor) standardized access to local file search via MCP or Agent Skills

When to avoid

  • you need an OSI-approved open-source license for commercial use - commercial deployment requires separate authorization
  • you need server-side or multi-user enterprise search across a network rather than a single-user desktop tool
  • your machine cannot run local AI models and you are unwilling to depend on cloud embedding/LLM APIs
  • you want a lightweight file indexer without downloading AI models or index-building overhead

Facets

application · maturity active

search-engine speech-recognition nlp machine-learning llm-inference mcp desktop-applications developer-tools files artificial-intelligence windows cross-platform python semantic-search local-search file-search multimodal-ai voice-search hybrid-search vector-search faiss whoosh ollama mcp-server agent-skills electron-app privacy-first chinese-language search natural-language-processing macos linux desktop

3 sources

Member repositories

RepositoryRoleHealth v2
dtsola/xiaoyaosearchmain70

For agents

markdown · JSON · MCP: product_card(name="dtsola/xiaoyaosearch")

Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem