Mcp Server Search
Give your coding agent live web search from the IDE without hand-copying results from a browser.
Overview
io.github.bytedance/mcp-server-search is an MCP server for the Idea phase that runs configurable web search queries for your agent.
What is this MCP server?
- stdio and SSE transports via npm package @agent-infra/mcp-server-search
- Configurable search engine (default: google) plus optional api-key and base-url for providers
- Designed for agent-infra web search operations inside MCP-compatible clients
- Fits early discovery workflows: competitors, APIs, pricing pages, and technical references
- Package identifier: @agent-infra/mcp-server-search (npm)
- Server schema version field: 1.0.0
- Transports: stdio and SSE (http://127.0.0.1:{port}/sse)
Community signal: 36.3k GitHub stars.
What problem does it solve?
Agents hallucinate or stale-date facts when you need current search results during research and scoping.
Who is it for?
Solo builders who want in-editor search during idea and validate research without building a custom search API wrapper.
Skip if: Teams that need full-site SEO crawls, PDF exports, or deep browser automation—that belongs to dedicated crawl or browser MCPs.
What do I get? / Deliverables
Your agent can issue real web searches from the MCP client and cite fresher sources while you explore opportunities.
- Agent-callable web search tool wired into your MCP config
- Configurable engine and optional provider endpoint settings
- Search results usable in downstream research and scoping prompts
Recommended MCP Servers
Journey fit
Solo builders validate ideas and scope features by searching competitors, docs, and market signals before they commit to a build path. Research is where agents need grounded, up-to-date URLs and snippets instead of guessing from training cutoffs.
How it compares
MCP search integration, not a brainstorming or planning skill.
Common Questions / FAQ
Who is io.github.bytedance/mcp-server-search for?
Indie and solo developers using Claude Code, Cursor, or other MCP clients who need live web search tools during research.
When should I use io.github.bytedance/mcp-server-search?
Use it in the Idea or Validate phases when you need competitors, docs, pricing, or API references that post-date model training.
How do I add io.github.bytedance/mcp-server-search to my agent?
Register @agent-infra/mcp-server-search from npm with stdio or SSE, set engine (default google), and add api-key or base-url if your provider requires them.