Retrieve
Answers are built from a finite, local corpus you can inspect, not from a model's memory.
Open source · MIT licensed · MCP-native
medground is an open-source retrieval-augmented generation (RAG) tool for biomedical literature. Ask a research question and get an answer backed by a real, verifiable citation, never a hallucinated one.
Built for researchers, evidence synthesis, and clinical teams.
MIT licensed. Python 3.11+. Runs locally.
Which treatment matches this patient's tumor?
A targeted therapy (osimertinib) that strongly slowed the cancer in trials.
Source: LAURA trial
Grounded. The source is real, and you can open it.
How it works
If a claim has no real source, it does not ship.
Answers are built from a finite, local corpus you can inspect, not from a model's memory.
Every claim carries a paper ID that links back to a real source.
A deterministic check confirms each citation exists before the answer reaches you.
The /doc skills
Follow one real patient case through six grounded skills. Hover the rings or the list to see what each one does, in plain language, with every answer backed by a real source.
The case
A lung cancer case
first treatment stopped working
A patient whose lung cancer returned after the first targeted treatment stopped working. Watch the case flow through six grounded skills, each answer backed by a real source.
a starting point, no claim yet
Six of sixteen skills shown. Type /doc to route by intent, or /doc-help for the full menu. Illustrative case, decision support, not medical advice.
# clone and install
git clone https://github.com/ArnaudTurn-pro/medgroundAI
cd medgroundAI
uv sync
# build a corpus
uv run medground ingest civic
# connect it to Claude
claude mcp add medground -- uv run --directory . medground-mcp
# optional: install the /doc slash commands
./install-skills.sh
Then ask Claude a grounded question, or type /doc to pick a skill. No tool names required.
medground surfaces what the published literature says. It is not clinically validated and makes no clinical decisions. A grounded citation means a source is real and retrievable, not that it is correct for a patient. That judgment is always a clinician's.