Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@small-mcp Count words in 'Hello world'"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A minimal FastMCP server in Python that demonstrates all four MCP primitives: tools, resources, resource templates, and prompts — including how to use the context object for logging and progress reporting.
uv (package manager)
Python 3.14+
Related MCP server: MCPServer (FastMCP)
uv syncuv run python main.pyThe server starts in stdio mode, ready to be connected to any MCP client.
Functions the LLM can call to perform actions.
Tool | Description |
| Add two numbers |
| Count words, characters, and lines |
| Return the current UTC time |
| Log a message via MCP context (demonstrates progress + logging) |
Static data the LLM can read.
URI | Description |
| Basic info about this server |
| A Python Hello World snippet |
Dynamic resources resolved from a URI pattern.
Template | Description |
| Hello World snippet for |
Reusable prompt templates the LLM can request by name.
Prompt | Args | Description |
|
| Ask for a summary in a given style |
|
| Request a code review |
|
| Explain a concept to a target audience |
The log_message tool shows how to inject ctx: Context into a tool to:
Send log messages back to the client via ctx.info()
Report progress via ctx.report_progress(progress, total)
uv run fastmcp dev inspector main.pyThen open the URL printed in the terminal to browse and call tools, read resources, and render prompts interactively.
Add this to your claude_desktop_config.json:
{
"mcpServers": {
"small-mcp": {
"command": "uv",
"args": ["run", "--directory", "/Users/nishantbhagat/Study/Projects/small mcp", "python", "main.py"]
}
}
}small mcp/
├── main.py # MCP server (all primitives in one file)
├── pyproject.toml # uv project config
└── uv.lockCall a tool by name with the given arguments.
Use this to execute tools discovered via search_tools.
Inputs: name (required), arguments.
Get a prompt by name with optional arguments.
Returns the rendered prompt as JSON with a messages array. Arguments should be provided as a dict mapping argument names to values.
Inputs: name (required), arguments. Returns: result
List all available prompts.
Returns JSON with prompt metadata including name, description, and optional arguments.
Inputs: no named parameters. Returns: result
List all available resources and resource templates.
Returns JSON with resource metadata. Static resources have a 'uri' field, while templates have a 'uri_template' field with placeholders like {name}.
Inputs: no named parameters. Returns: result
Read a resource by its URI.
For static resources, provide the exact URI. For templated resources, provide the URI with template parameters filled in.
Returns the resource content as a string. Binary content is base64-encoded.
Inputs: uri (required). Returns: result
Search for tools using natural language.
Returns matching tool definitions ranked by relevance, in the same format as list_tools.
Inputs: query (required). Returns: result
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 6 tools
Each tool has a clear, distinct role: search vs call, list vs read, list vs get. There is no overlap in purpose.
All tool names follow a consistent verb_noun pattern (search_tools, call_tool, list_resources, read_resource, list_prompts, get_prompt).
With 6 tools covering three resource types (tools, resources, prompts), the count is well-scoped and appropriate for a small MCP utility server.
The server covers tools, resources, and prompts, but misses a list_tools tool, which is a notable gap since search_tools cannot enumerate all tools. This could cause agent failures when trying to discover the full tool surface.
A simple MCP server built with FastMCP and python
Nifty's MCP server — exposes tasks, projects, messages, and files as tools for AI agents.
MCP server for progressive tool usage at any scale (see https://klavis.ai)
