learning-mcp-server
Enables a local Ollama model to drive tool selection and execution, allowing the AI to decide which tools to call and process results for a final answer.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
learning-mcp
A minimal MCP server and client, built on the official mcp Python SDK, exercising both
supported transports: stdio and streamable HTTP.
The server exposes:
Tools:
add(a, b),echo(text)Resource:
memo://notesPrompt:
summarize(text)
Setup
uv syncRelated MCP server: node-mcp-poc
Run it
With ollama serve running and the default model (gemma4) pulled, just run:
uv run learning-mcp-clientThis spawns the server over stdio and lets a local Ollama model decide which tools to call: it lists the server's tools, hands their schemas to Ollama, executes whatever tool calls the model requests, and feeds the results back until Ollama gives a final answer.
Override the model or prompt:
uv run learning-mcp-client --prompt "Echo back 'hello'" --model qwen3.5:2bRun the built-in tool/resource/prompt demo instead (no LLM involved):
uv run learning-mcp-client --demosrc/learning_mcp/tool.py defines a local tool and an MCP tool in one file
and shows what actually differs at registration: the MCP tool (add) is
registered up front with @server.tool(); the local tool (multiply) is
never registered at all, it's just a function handed to Ollama at call
time. See "Local tool vs. MCP tool" in MCP_Overview.md. Requires ollama serve running with gemma4 pulled:
uv run learning-mcp-local-vs-mcpRun over HTTP
Start the server in one terminal:
uv run learning-mcp-server --transport http --host 127.0.0.1 --port 8000Then, in another terminal:
uv run learning-mcp-client --transport http --url http://127.0.0.1:8000/mcpTool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Remote MCP server exposing SMI Aware tools, resources, and skills over Streamable HTTP.
AI Reasoning Cache & Consensus Layer with 11 MCP tools via Streamable HTTP.
Nifty's MCP server — exposes tasks, projects, messages, and files as tools for AI agents.
MCP server for your apps' tools and custom tools, plus hosted AI agents and approval-gated workflows
Related MCP Servers
- FlicenseNot gradedqualityDmaintenanceProvides math and weather tools accessible via LangGraph agent using MCP protocol with stdio and streamable HTTP transports.1-
- FlicenseNot gradedqualityDmaintenanceA minimal MCP server that exposes tools for addition, echoing text, time lookup, and URL fetching, with support for HTTP and stdio transports.-
- FlicenseNot gradedqualityBmaintenanceEnables MCP-compatible AI clients to invoke CLI-driven agent tools over Streamable HTTP, including shell execution, file operations, patching, image viewing, web search, and nested agent tasks, with permission modes and real-time progress streaming.-
- AlicenseAqualityCmaintenanceProvides a Model Context Protocol server with tools for knowledge base search, retrieval, and addition, plus system time, resources, and prompts. Supports stdio and Streamable HTTP transports for local and shared deployments.4MIT