MCP研究助手

Option 2: One command (if you have Go installed)

go install github.com/zoharbabin/web-researcher-mcp/cmd/web-researcher-mcp@latest

Option 3: Docker

docker run -e GOOGLE_CUSTOM_SEARCH_API_KEY=YOUR_KEY \
           -e GOOGLE_CUSTOM_SEARCH_ID=YOUR_CX \
           docker.io/zoharbabin/web-researcher-mcp:latest

<details> <summary><strong>Option 4: Build from source</strong></summary>

git clone https://github.com/zoharbabin/web-researcher-mcp.git
cd web-researcher-mcp
go build -o web-researcher-mcp ./cmd/web-researcher-mcp

</details>

Setup for Each AI App

Quick Start

Option 1: Download (simplest)

Download the ready-to-use binary for your system from Releases. No programming tools needed.

Configuration

You need at least one search provider's API key. Pick whichever is easiest for you:

Option A: Google (default)

Option B: Brave Search (simpler signup)

Set SEARCH_PROVIDER=brave and you're done. No Google keys needed.

Tip: You can set up multiple providers for automatic backup — see Search Providers below.

<details> <summary><strong>All Search Provider Options</strong></summary>

</details>

Academic Search (Optional — no signup needed)

With these set, academic_search returns real papers with DOIs, authors, citation counts, and open-access PDF links. Without them, it still works but uses web search as a fallback.

Patent Search (Optional)

With these, patent_search returns structured patent data with classification codes, dates, and inventors. Without them, it falls back to web search.

<details> <summary><strong>Advanced: HTTP mode, OAuth, and all other settings</strong></summary>

See docs/DEPLOYMENT.md for the complete list of all settings (cache, rate limiting, scraping, observability, etc.).

</details>

---

How It Compares

Troubleshooting

<details> <summary><strong>Server starts but tools fail with "API key" errors</strong></summary>

The server starts even with missing credentials (to allow MCP handshake). Set your API keys in the env block of your MCP client config, not in your shell profile.

</details>

<details> <summary><strong>Some pages come back empty</strong></summary>

For JavaScript-heavy sites, the tool uses a real browser (Chromium). It auto-downloads on first use (~200MB). If you already have Chrome installed, set CHROME_PATH to point to it, or use the Docker image which has everything included.

</details>

<details> <summary><strong>Cache serving stale results after upgrade</strong></summary>

The disk cache lives at your OS cache directory (e.g., ~/Library/Caches/web-researcher-mcp/ on macOS, ~/.cache/web-researcher-mcp/ on Linux). Delete that directory to clear it, or set CACHE_DIR to a custom path.

</details>

<details> <summary><strong>Hitting search limits (429 errors)</strong></summary>

Google's free tier allows 100 searches/day. If you're hitting that: - Switch to Brave Search (SEARCH_PROVIDER=brave) — more generous free tier - Set up multiple providers (SEARCH_ROUTING=brave,google) — if one is rate-limited, it uses the other - Or upgrade Google to paid ($5 per 1,000 searches)

</details>

---