FieldCure.Mcp.Essentials

FieldCure.Mcp.Essentials

Install once, get the basics. An MCP server with 20 essential tools for any MCP client — web search with runtime engine switching (get/set), Wolfram|Alpha, web/document fetching, URL file downloads, shell, JavaScript sandbox, file I/O, and persistent memory. Category search (news, images, scholar, patents) is always available with per-tool runtime capability guards against the active engine.

Tools

Tool	Description
http_request	Full HTTP client (GET/POST/PUT/DELETE/PATCH/HEAD) with SSRF protection and max_response_chars for response size control
web_search	Search the web and return snippets (title, URL, description)
web_fetch	Fetch a URL and extract content as Markdown — HTML pages and documents (PDF, DOCX, HWPX, PPTX, XLSX)
download_file	Download URL content to disk with a configurable download directory, 100 MB limit, and atomic save
run_command	Shell command execution with timeout, working directory, env vars, explicit shell selection, and truncation flags
run_javascript	Sandboxed JavaScript (Jint) — math, JSON, regex, data processing
wolfram_alpha	Wolfram|Alpha computational knowledge — symbolic math, plots, unit conversions, constants. MathML passes through for native rendering
get_environment	System info — time, timezone, OS, hostname, username
read_file	Read files — text with offset/limit, documents (PDF, DOCX, HWPX, PPTX, XLSX) parsed to Markdown
write_file	File writing (overwrite/append) with auto directory creation
search_files	File search by glob pattern and content (grep-like)
remember	Store a key-value memory (persisted in SQLite)
forget	Delete memories by key or keyword search
list_memories	Search and list stored memories with FTS5 and pagination
set_search_engine	Switch the active search engine (bing, duckduckgo, serper, tavily, serpapi) at runtime. Paid-engine API keys resolve lazily via env var or MCP Elicitation on the next search. Emits notifications/tools/list_changed on success.
get_search_engine	Return the currently active engine and its category capabilities. Read-only; useful for host UIs reflecting live engine state or for confirming category support before calling a category search tool.

Category Search (SerpApi / Serper / Tavily)

Always registered. Each tool runtime-guards on the active engine's capabilities and returns a descriptive error pointing at set_search_engine when the current engine does not support the category.

Tool	Description	SerpApi	Serper	Tavily
search_news	Search recent news articles via Google News	Yes	Yes	Yes
search_images	Search images with size/type filtering	Yes	Yes	—
search_scholar	Search academic papers with citation counts	Yes	Yes	—
search_patents	Search patent documents with inventor/assignee filtering	Yes	Yes	—

Document Parsing

web_fetch and read_file can parse binary documents into Markdown:

Format	Extension	Detection
PDF	.pdf	Content-Type / URL extension (text-layer only; scanned pages yield empty text — use fieldcure-mcp-rag for OCR)
Word	.docx	Content-Type / URL extension
Hangul (HWPX)	.hwpx	URL extension (no standard Content-Type)
PowerPoint	.pptx	Content-Type / URL extension
Excel	.xlsx	Content-Type / URL extension

Output includes headings, tables, math expressions ([math: LaTeX]), and slide/page separators.

File Downloads

download_file saves the original bytes from an HTTP(S) URL. If save_path is omitted, the tool infers a filename from Content-Disposition, the URL path, or a generated fallback name. Relative save_path values resolve under the configured download_directory; absolute paths are used as-is except for protected system directories.

The default download directory is ~/Downloads/mcp and is created automatically on first use. Downloads are written to a temporary file in the destination directory and then committed with an atomic move/replace, so failed or cancelled downloads do not leave a partial final file.

Configure it in %LOCALAPPDATA%/FieldCure/Mcp.Essentials/settings.json:

{
  "download_directory": "~/Downloads/mcp"
}

Override with --download-directory <path> or ESSENTIALS_DOWNLOAD_DIRECTORY.

Web Search

Default engine is Bing (free, no API key needed). For more reliable results, use an API-based engine:

Engine	Free Tier	API Key
Bing (default)	Unlimited (scraping)	Not needed
Serper	2,500 one-time	serper.dev
Tavily	1,000/month	tavily.com
SerpApi	100/month	serpapi.com

# Use Serper
fieldcure-mcp-essentials --search-engine serper --search-api-key YOUR_KEY

# Or via environment variables
ESSENTIALS_SEARCH_ENGINE=tavily ESSENTIALS_SEARCH_API_KEY=xxx fieldcure-mcp-essentials

Engine-specific environment variables are also auto-detected when --search-engine is omitted:

Engine	Environment Variable
Serper	SERPER_API_KEY
SerpApi	SERPAPI_API_KEY
Tavily	TAVILY_API_KEY

Use the region parameter for localized results:

{ "query": "서울 맛집", "region": "ko-kr" }

Without --search-engine, a fallback engine (Bing → DuckDuckGo) auto-switches on CAPTCHA.

If a paid engine is selected explicitly but no API key is configured, the server no longer silently falls back. On the first web_search call it asks the MCP client for the key via MCP Elicitation; if the client does not support Elicitation, or the user declines and then also declines the follow-up "use free fallback?" prompt, the call soft-fails with a clear message. Cached keys live for the process lifetime.

Wolfram|Alpha

wolfram_alpha calls the Full Results API v2 and returns mixed text/image content. MathML is passed through verbatim — clients that render MathML (e.g. ChatPanel WebView2) display formulas natively, no client-side conversion needed. Plots and other visual pods are embedded as ImageContent. The API's reinterpret=true flag is always on, so the server auto-corrects most failed queries before returning.

Set WOLFRAM_APPID to the AppID obtained at developer.wolframalpha.com (select Full Results API, free tier: 2,000 calls/month, non-commercial). Clients that support MCP Elicitation are prompted for the key on first use if the env var is unset.

⚠️ Use developer.wolframalpha.com, not developer.wolfram.com — the latter is a separate paid portal.

The tool is always registered regardless of AppID status; without a key it returns a setup-guidance error so the model can inform the user.

Run Command

run_command defaults to cmd.exe on Windows and /bin/sh on Unix for backward compatibility. Pass shell to opt into pwsh, powershell, cmd, bash, or sh. Responses include shell_used, stdout_truncated, and stderr_truncated; max_output_chars caps stdout and stderr independently.

Memory

Memories are stored in SQLite (%LOCALAPPDATA%/FieldCure/Mcp.Essentials/memory.db) and shared across all MCP clients on the same machine.

# Custom memory path
fieldcure-mcp-essentials --memory-path /path/to/memory.db

# Or via environment variable
ESSENTIALS_MEMORY_PATH=/path/to/memory.db fieldcure-mcp-essentials

Quick Start

dotnet tool install -g FieldCure.Mcp.Essentials

Claude Desktop

{
  "mcpServers": {
    "essentials": {
      "command": "fieldcure-mcp-essentials"
    }
  }
}

VS Code

{
  "servers": {
    "essentials": {
      "command": "fieldcure-mcp-essentials"
    }
  }
}

Requirements

• .NET 8.0 Runtime

Optional

Works out of the box with no additional dependencies. If available on your system, run_command can leverage tools like Node.js, Python, Git, Docker, and any other CLI tools.

See Also

Part of the AssistStudio ecosystem.

Links

• GitHub
• License: MIT