About
Simple experiments with MCPs and agentic frameworks
README
Pydantic AI and LangGraph-based agent wired to an MCP server for working with a local Zettelkasten notes corpus and reminders (Things.app).
Uses uv for environment, dotenv for config, and a local OpenAI/Ollama-compatible runtime by default.
Requirements
- Python 3.14+
- uv installed (https://docs.astral.sh/uv/)
- External tools on PATH for notes operations:
- get-notes-by-level (from https://github.com/romanthekat/r-notes)
- relevant-notes (from the same repo)
- grep, sed
- macOS + Things app installed for reminders (optional; required for add_reminder)
- Local LLM runtime exposing an OpenAI-compatible API (default base_url http://localhost:1234/v1)
Setup
- Sync deps
uv sync
- Configure environment
cp .env.example .envand set at minimum:- NOTES_PATH=/absolute/path/to/your/markdown/notes
- Optional/typical:
- OPENAI_API_KEY (if your runtime checks it)
- LLM_MODEL (default: openai/gpt-oss-120b)
- LLM_BASE_URL (default: http://localhost:1234/v1)
- LLM_TEMPERATURE (default: 1)
- AGENT_RECURSION_LIMIT (default: 42)
- AGENT_THREAD_ID (default: some thread id)
Running
- Start the interactive agent:
uv run main.py
- MCP transport defaults to stdio, spawning the server via
uv run mcp_server.pyunder the hood.- To experiment with HTTP, see commented example in main.py and run: uv run mcp_server.py
Notes
- NOTES_PATH must be an absolute, existing directory. The server will raise a clear error if misconfigured.
- Ensure your shell PATH (where the external CLIs live) is visible to processes launched via uv.
Tests (optional)
- A trivial smoke test can be run without external services:
- uv run python -m unittest -v test_smoke.py
Pydantic-AI Versions
In addition to the LangGraph main.py, there are versions implemented using Pydantic-AI:
main_pydantic.py: Single agent setup with MCP tools.main_pydantic_moe.py: Multi-agent "Mixture of Experts" setup where a Facilitator orchestrates a discussion between an Analyst, Strategist, and Critic.
Installing Experiments Ml
This server has no published package — it is built from source. Open the repository and follow its README.
▸ github.com/romanthekat/experiments-mlFAQ
Is Experiments Ml MCP free?
Yes, Experiments Ml MCP is free — one-click install via Unyly at no cost.
Does Experiments Ml need an API key?
No, Experiments Ml runs without API keys or environment variables.
Is Experiments Ml hosted or self-hosted?
Self-hosted: the server runs locally on your machine via the install command above.
How do I install Experiments Ml in Claude Desktop, Claude Code or Cursor?
Open Experiments Ml on unyly.org, pick your client tab (Claude Desktop, Claude Code, Cursor) and press Install — the config is generated automatically, no JSON editing.
Related MCPs
GitHub
PRs, issues, code search, CI status
by GitHubFilesystem
Secure file operations with configurable access controls.
Memory
Knowledge graph-based persistent memory system.
Template MCP Server
A CLI tool to create a new Model Context Protocol server project with TypeScript support, dual transport options, and an extensible structure
by mcpdotdirectAmap Maps Mcp Server
MCP server for using the AMap Maps API
by duxiaohuiSupabase
Database, auth and storage
by SupabaseEverything
Reference / test server with prompts, resources, and tools.
Git
Tools to read, search, and manipulate Git repositories.
Sequential Thinking
Dynamic and reflective problem-solving through thought sequences.
Time
Time and timezone conversion capabilities.
Compare Experiments Ml with
Not sure what to pick?
Find your stack in 60 seconds
Author?
Embed badge for your README
Browse similar
All development MCPs
