Skip to main content
Glama
Sign In
CubicJ MCP Search

CubicJ MCP Search

Official
by cubicj
OverviewSchemaSourceRelated ServersDiscussionsAdmin
TypeScript
Remote

CubicJ MCP Search

Stateless Streamable HTTP MCP server with two read-only tools for LLM clients:

Tool

What it does

web_search

Brave LLM Context search: relevance-ranked page excerpts per source URL, with count, max_tokens, freshness, country, and search_lang controls

fetch

Readable extraction of one URL as markdown (Defuddle), with site extractors for DCInside, Naver Blog, Bluesky, GitHub, Clien, Ruliweb, FMKorea, and namu.wiki, PDF text, raw mode, and pagination

It replaces the separate brave-search-api-mcp, dc-mcp-search, and defuddle-fetch-mcp servers with one deployment sized by measured usage.

Requirements

  • Node.js 24, npm 11

  • BRAVE_API_KEY (Brave Search API, Search plan) for web_search

  • Optional GITHUB_TOKEN for higher GitHub API limits in fetch

Related MCP server: Brave Search MCP Server

Run

npm ci
npm run build
BRAVE_API_KEY=... node dist/index.js --transport http --port 8006
curl -s http://127.0.0.1:8006/health

Without --transport http (or --port) the server speaks MCP over stdio. PORT sets the HTTP port when --port is absent; .env.example lists the variables for a systemd EnvironmentFile.

The HTTP endpoint is POST /mcp. Requests carrying a browser Origin header are rejected, bodies are capped at 1 MiB, and one log line per request goes to stderr. Put authentication in front of it (a reverse proxy or an OAuth gateway); the server has none.

Development

npm test            # builds, then runs Vitest offline
npm run typecheck
npm run lint        # Biome
npm run smoke -- "query"   # live web_search, needs BRAVE_API_KEY exported

License

MIT

Available Tools

2 tools
fetchFetch URLA
Read-onlyIdempotent

Fetch one URL and return its readable content as markdown with a title and metadata. Use it to read a page in full after web_search, or any URL the user gives: articles, docs, forums, news. Site-specific extractors return post body plus comments for gall.dcinside.com, Clien, Ruliweb, and FMKorea; body only for blog.naver.com and namu.wiki; threads with replies and profiles with recent posts for bsky.app; raw content or REST data for github.com files, repos, issues, pull requests, and gists. Non-HTML text (JSON, XML, RSS, plain text) is returned as-is; PDFs as page-separated text with page_count; image-only PDFs return a notice; other binaries are errors. raw=true skips extraction and returns the decoded response text. Long content is paginated: content is sliced from start_index (default 0) for max_length characters (default 50000); to continue, call again with start_index advanced by the previous max_length and stop when start_index + max_length >= total_length. Bot-protection pages return an error; empty pages that need JavaScript or a login return a notice.

Inputs: url (required), raw, max_length, start_index. Returns: url, mode, title, content, +5 more

web_searchWeb SearchA
Read-onlyIdempotent

Search the web and get relevance-ranked page excerpts for each result, ready to read and cite. First tool for any research, fact lookup, or URL discovery; most questions need one call and no follow-up fetch. Operators use a colon: site:example.com, "exact phrase", -excluded. count sets how many source URLs (1-20, default 8); max_tokens caps the total excerpt size (1024-32768, default 8192); freshness limits results by date (pd, pw, pm, py, or YYYY-MM-DDtoYYYY-MM-DD); country and search_lang scope the market. Returns JSON: query, result_count, results[{title, url, site, age, snippets[]}].

Inputs: query (required), count, country, freshness, +2 more. Returns: query, results, result_count

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    A4.4/5.0

    Scored across 2 tools

    Disambiguation5/5

    web_search handles discovery with ranked excerpts, while fetch retrieves and extracts full content from a single URL. Their roles are clearly distinct and the descriptions reinforce when to use each.

    Naming Consistency3/5

    fetch is a bare verb and web_search is a compound noun/verb form, so there is no shared verb_noun pattern. Both are readable and common, but the conventions are mixed.

    Tool Count4/5

    Two tools cover the core search-and-retrieve workflow without redundancy, but the count is slightly below the typical 3–15 range. Each tool earns its place for the server's narrow purpose.

    Completeness4/5

    The surface covers discovery via web_search and retrieval via fetch, including pagination and many site-specific extractors. Minor gaps like batch fetching or search pagination exist, but there are no major dead ends.

    Maintenance

    ActivityMaintained
    ResponsivenessOutdated data

    Resources

    Related MCP Connectors

    Related MCP Servers

    • Brave Search MCP Server

      arben-adm
      A
      license
      B
      quality
      D
      maintenance
      Facilitates integration of Brave Search functionalities with AI assistants through the Model Context Protocol, enabling web and local searches using the Brave Search API.
      2
      15
      MIT

    • Brave Search MCP Server

      YC-Agents-Hackathon
      A
      license
      Not graded
      quality
      D
      maintenance
      Enables web and local business searches through the Brave Search API. Provides general web search with pagination and filtering, plus local business search with automatic fallback to web results.
      MIT

    • search-online-mcp

      gastonmorixe
      A
      license
      Not graded
      quality
      D
      maintenance
      Exposes a standardized web search tool via Brave Search for MCP clients like Codex, Claude, and Cursor, returning stable JSON.
      MIT