Fastmcp Research Agent
No description available
Ask AI about Fastmcp Research Agent
Powered by Claude Β· Grounded in docs
I know everything about Fastmcp Research Agent. Ask me about installation, configuration, usage, or troubleshooting.
0/500
Reviews
Documentation
FastMCP Γ PydanticAI Research Agent
π§ How to Run
- Create a virtual environment:
python3 -m venv venv source venv/bin/activate pip install -U pip pip install -r requirements.txt
python -m student.run --topic "US semiconductor export controls (2024β2025) overview" --out artifacts
π How to Run This Project
Follow these steps to set up and run the FastMCP Γ PydanticAI Research Agent locally:
1οΈβ£ Clone this repository git clone https://github.com/siddharth1956/fastmcp-research-agent.git cd fastmcp-research-agent
2οΈβ£ Create a virtual environment python3 -m venv venv source venv/bin/activate # macOS / Linux
or on Windows:
venv\Scripts\activate
3οΈβ£ Install dependencies pip install -U pip pip install -r requirements.txt
4οΈβ£ Create your .env file Copy the example file and update it with your own OpenAI API key: cp .env.example .env
Open .env and fill in the values: OPENAI_API_KEY=sk-yourkeyhere OPENAI_MODEL=openai:gpt-4o-mini MCP_HTTP_URL=http://127.0.0.1:8765 ASSIGNMENT_MODE=dev
β οΈ Do not commit .env to GitHub. It contains your private API key.
5οΈβ£ Run the research agent python -m student.run --topic "US semiconductor export controls (2024β2025) overview" --out artifacts
This command: β’ Connects to the MCP server. β’ Runs the PydanticAI agent. β’ Produces a research brief and logs in the artifacts/ folder.
βΈ»
6οΈβ£ Check generated outputs ls -la artifacts cat artifacts/brief.md cat artifacts/report.json sqlite3 notes.sqlite "SELECT COUNT(*) FROM notes;" You should see: β’ brief.md β 5β8 paragraph synthesis with citations [1], [2] β’ report.json β includes "pii_flagged": false β’ search_results.json, selected_urls.json, snippets.json β’ notes.sqlite β persistent memory database
βΈ»
7οΈβ£ Optional (for graders)
If youβre grading or running in a sandboxed environment (no API key): export ASSIGNMENT_MODE=grading python -m student.run --topic "Any test topic" --out artifacts The project will run with stubbed MCP responses β no internet or API key required.
βΈ»
β Folder Structure fastmcp-research-agent/ ββ student/ β ββ agent.py β ββ run.py β ββ mcp_client.py β ββ init.py ββ artifacts/ β ββ .gitkeep ββ .env.example ββ .gitignore ββ README.md ββ requirements.txt
