Fastmcp Client CLI

FastMCP CLI: List and Call

Use fastmcp discover, fastmcp list, and fastmcp call to interact with MCP servers from the command line. These commands are FastMCP 3.x client commands and are useful for development, debugging, scripting, and shell-capable agents that do not have native MCP support.

Listing Tools

# Remote server
fastmcp list http://localhost:8000/mcp

# Local Python file (runs via fastmcp run automatically)
fastmcp list server.py

# MCPConfig with multiple servers
fastmcp list mcp.json

# Stdio command (npx, uvx, etc.)
fastmcp list --command 'npx -y @modelcontextprotocol/server-github'

# Include full input/output schemas
fastmcp list server.py --input-schema --output-schema

# Machine-readable JSON
fastmcp list server.py --json

# Include resources and prompts
fastmcp list server.py --resources --prompts

Default output shows tool signatures and descriptions. Use --input-schema or --output-schema to include full JSON schemas, --json for structured output. With --json, resources and prompts are included only when --resources or --prompts is also passed.

Calling Tools

# Key=value arguments (auto-coerced to correct types)
fastmcp call server.py greet name=World
fastmcp call server.py add a=3 b=4

# Single JSON object for complex/nested args
fastmcp call server.py create_item '{"name": "Widget", "tags": ["a", "b"]}'

# --input-json with key=value overrides
fastmcp call server.py search --input-json '{"query": "hello", "limit": 5}' limit=10

# JSON output for scripting
fastmcp call server.py add a=3 b=4 --json

Type coercion is schema-driven: integers, numbers, booleans, arrays, and objects are coerced from strings based on the target tool's input schema. Booleans accept true/false, yes/no, and 1/0. Arrays and objects must be valid JSON.

fastmcp call can also read a resource URI or get a prompt:

# Resource URI targets are detected by ://
fastmcp call server.py resource://docs/readme

# Prompt targets require --prompt
fastmcp call server.py summarize --prompt topic=FastMCP

Server Targets

All commands accept the same server targets:

Python file targets are launched through fastmcp run <file> --no-banner; code in an if __name__ == "__main__" block is not used by fastmcp run. JSON targets for list and call must be MCPConfig-style files with an mcpServers key. A fastmcp.json server config is not a direct client target; start it first with fastmcp run.

Servers configured in editor configs (Claude Desktop, Claude Code, Cursor, Goose) or the current directory's ./mcp.json can be referenced by name. Bare names are rejected if the same name appears in multiple sources. Use source:name (e.g. claude-code:my-server, cursor:weather) to target a specific source. Run fastmcp discover to see available names.

For SSE servers, pass --transport sse:

fastmcp list http://localhost:8000/mcp --transport sse

HTTP URLs default to Streamable HTTP. The --transport sse override is for older SSE-only servers.

Auth

HTTP targets automatically use OAuth unless disabled. If the server does not require auth, OAuth setup is a no-op. Disable auth or pass a bearer token with --auth:

fastmcp call http://server/mcp tool --auth none
fastmcp list http://server/mcp --auth "Bearer sk-..."

Top-level auth is not applied to MCPConfig JSON targets; those transports are resolved per configured server.

Discovering Configured Servers

# See all MCP servers in editor/project configs
fastmcp discover

# Filter by source
fastmcp discover --source claude-code

# JSON output
fastmcp discover --json

Scans Claude Desktop, Claude Code, Cursor, Goose, and ./mcp.json. Sources: claude-desktop, claude-code, cursor, goose, project. Claude Code discovery includes global ~/.claude.json servers and project-scoped entries matching the current working directory.

Workflow Pattern

Discover tools first, then call them:

# 1. See what servers are configured
fastmcp discover

# 2. See what tools a server has
fastmcp list weather

# 3. Call a tool
fastmcp call weather get_forecast city=London

If you call a nonexistent tool, FastMCP suggests close matches.

For shell-agent integration, prefer fastmcp list --json to gather tool schemas and fastmcp call --json to parse content blocks, is_error, and structured content reliably.