MEVA Health AI MCP Server
Click on "Install 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., "@MEVA Health AI MCP ServerVerify the medical claim that patient 001 has hypertension."
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.
MEVA
Medical Evidence Verification Agent
What it does: MEVA evaluates whether AI-generated medical-record claims are grounded in retrieved synthetic FHIR evidence, using a deterministic (non-AI) verifier โ no diagnosis, no treatment advice, no real patient data.
๐งช Live Sandbox
Try MEVA in your browser โ no Ollama, no API key, no installation required.
โถ Open Live Sandbox
Synthetic data only โ 21 fictional, Synthea-generated patients
Deterministic verifier โ every result comes from plain Python evidence matching, not a model's opinion
No AI model runs in the public sandbox โ you build the claim yourself and MEVA checks it against real recorded (synthetic) data
Try Live Sandbox ยท Quick Start ยท Benchmark Methodology ยท Contribute
MEVA is not a medical chatbot, diagnostic AI, clinical decision support tool, treatment recommendation system, or medical device. It is not clinically validated. All patient data is 100% synthetic. See Scope and safety below and
docs/safety-and-scope.mdfor the full statement.
Related MCP server: MediLinkAI
See MEVA in action
Screenshots from the live hosted sandbox โ synthetic data only, not medical advice, and no AI model runs in the hosted sandbox (every result comes from MEVA's deterministic verifier, not a model's opinion).
Guided Mode

Guided Mode โ choose a fictional patient and verify a claim using a simple plain-English workflow.
Evidence-backed verification

MEVA checks the claim against recorded synthetic FHIR evidence and returns SUPPORTED, CONTRADICTED, UNSUPPORTED, or UNVERIFIABLE.
Advanced Mode

Advanced Mode โ technical evidence explorer and structured claim controls for developers, researchers, and contributors.
Why MEVA exists
Local AI agents can call tools, retrieve real data, and generate a structured "answer" โ but nothing forces that structured answer to actually match the evidence the agent retrieved. MEVA measures that gap directly, with a verifier that never trusts a model's self-report of its own correctness.
Architecture
flowchart TD
A[Synthetic Synthea FHIR data] --> B[MEVA FHIR layer]
B --> C[MCP tools]
C --> D[Local AI model - via Ollama]
D --> E[Natural-language answer]
E --> F[Claim extraction]
F --> G[Deterministic evidence verifier]
G --> H[Benchmark / report]MEVA supports two evaluation modes, reported separately and never
combined into one score (see docs/decoupled-evaluation.md):
END_TO_END โ the tested model answers a question AND encodes its own answer into MEVA's structured
MedicalClaimschema, in one pass.DECOUPLED โ the tested model answers in prose only; a separate fixed extractor model converts that saved prose into structured claims, which are verified the same way. This isolates "did the model know the right answer" from "did the model correctly format JSON."
In both modes, the final verification step โ matching a claim against real evidence โ is always plain, deterministic Python. No LLM ever judges its own (or another model's) correctness.
Quick start
git clone https://github.com/Tanz2024/meva-health-ai
cd meva-health-ai
python3 -m venv .venv
source .venv/bin/activate
pip install -e .
pytestMEVA is designed to be run from within a cloned copy of this repository (as above โ
git clonethen an editable install), not as a standalone package installed from elsewhere. Its synthetic FHIR fixtures (data/synthetic/synthea/) and benchmark definitions (benchmarks/) are read from repository-relative paths, not bundled as installable package data โ apip installof a built wheel/sdist outside a repo checkout will not have patient data available. This is current, intentional scope (a research/engineering repo, not a distributed library) โ seedocs/publishing-checklist.mdif this changes in the future.
This runs the entire offline test suite (no AI model required โ see What runs without AI).
Optional โ local AI:
ollama pull qwen3:4b
python3 examples/verify_local.pySee examples/ for more runnable scripts, and docs/local-ai.md for how
MEVA talks to Ollama.
What runs without AI (Ollama not required)
Most of MEVA works with no AI model at all:
FHIR parsing (
src/meva/fhir/)Deterministic evidence verification (
src/meva/verification/)Benchmark dataset loading and validation (
meva.benchmark.validator)The full offline test suite (
pytest)The verifier-challenge examples (
examples/verify_contradiction_demo.py) โ these test MEVA's own verification logic with a hand-written wrong claim, no live model involved
Ollama is only needed when actually running local model inference
(examples/ask_local.py, examples/chat_local.py) or model-assisted claim
extraction (examples/run_decoupled_pilot.py, run_decoupled_full.py,
run_extractor_fidelity.py).
Synthetic data
MEVA's public patient fixtures (data/synthetic/synthea/patient-01.json
through patient-21.json) are generated locally, by this project, using
the official Apache-2.0-licensed Synthea
generator (pinned to tag v3.4.0), with a fixed, documented, reproducible
seed. No real patient data is included anywhere. Full generation details
(exact command, seed, and per-file SHA-256 hashes) are in
data/synthetic/synthea/PROVENANCE.md.
See docs/synthetic-data.md for the full picture,
including why an earlier set of 18 patients (used through Stage 8A) was
replaced โ that earlier set had been copied from a repository with no
declared license and is no longer part of the public dataset (see
docs/historical-sample-data-provenance.md).
Benchmark results
MEVA has two benchmark datasets on record, and they must not be conflated:
PUBLIC REPRODUCIBLE DATASET: benchmark v0.4 โ built entirely from the
locally-generated Apache-2.0 fixtures above (53 cases, 16 unique patients;
see benchmarks/v0.4/manifest.json).
v0.4 model-comparison results are pending โ no qwen3:4b/llama3.2:3b run
has been performed against v0.4 yet.
HISTORICAL DEVELOPMENT RESULT: benchmark v0.3 โ full qwen3:4b vs
llama3.2:3b results below. This was measured against the now-removed
former patient set (see above) โ it remains a valid historical development
record of the methodology and findings, but is not a result on the
current public v0.4 dataset, and the two must not be compared directly.
Full report: docs/baseline-results-v0.3.md
(the numbers below use the corrected Stage 7C2.1 verifiable-coverage
formula โ see that document for the correction history; the original,
uncorrected numbers are also disclosed there, not hidden).
No winner is declared. Read grounding score together with verifiable coverage โ a high grounding score computed over very few checkable claims looks better than it is.
Retrieval + END_TO_END structured-output metrics (v0.3, historical)
qwen3:4b | llama3.2:3b | |
Tool recall | 1.000 | 0.981 |
Tool precision | 1.000 | 1.000 |
Exact tool match | 1.000 | 0.962 |
Evidence recall | 0.810 | 0.738 |
E2E structured validity | 0.917 | 0.087 |
E2E verifiable coverage | 0.656 | 0.120 |
E2E grounding | 83% | 70% |
DECOUPLED evaluation (v0.3, historical; separate, fixed qwen3:4b extractor)
qwen3:4b | llama3.2:3b | |
DECOUPLED verifiable coverage | 0.990 | 0.987 |
DECOUPLED grounding | 89% | 80% |
DECOUPLED evaluation uses qwen3:4b as a fixed claim extractor for both
models' saved answers, including qwen3:4b's own โ this introduces potential
extractor-specific bias, documented explicitly in
docs/decoupled-evaluation.md. END_TO_END and DECOUPLED answer different
questions and must never be read as "the model got better."
Extractor validation โ the extractor is not perfect
Development (10 fixtures) | Holdout (14 unseen fixtures) | |
Precision | 1.000 | 0.929 |
Recall | 1.000 | 0.813 |
F1 | 1.000 | 0.867 |
Exact claim-set match | 1.000 | 0.857 |
Negative-claim preservation | 1.000 | 1.000 |
Attribute accuracy | 1.000 | 0.750 |
Do not read the ~99% DECOUPLED coverage numbers above as "99% extraction accuracy." Coverage measures how much of the extractor's output the verifier could check; the holdout numbers here measure whether that output actually matches what the source answer said.
Observation-category finding
Both models scored unusually low on observation-category questions:
qwen3:4b 20%, llama3.2:3b 0% grounding (n=10 cases each). Stage 7D2.2
independently audited every observation case against MEVA's real FHIR data
and tool layer and found no infrastructure or evaluation bug that
invalidates these results โ 18 of 20 model-case pairs are genuine model
grounding errors. Full audit: docs/observation-audit.md. This describes
benchmark behavior, not clinical performance.
Try MEVA (deterministic verification only, no AI model)
Four ways to explore MEVA's deterministic verifier against the 21 public v0.4 synthetic patients โ none of them require an AI model:
Public hosted sandbox (no install): Open Live Sandbox
Local browser sandbox:
pip install -e ".[playground]"
streamlit run streamlit_app.pyLocal CLI playground:
python3 examples/playground.py demo
python3 examples/playground.py list-patients
python3 examples/playground.py verify --patient-id <id> --category allergy --assertion present --value "Peanut"Full local AI mode (optional, needs Ollama โ see What runs without AI below).
All four share the same service layer (meva.playground) and call MEVA's
real, unmodified verifier โ you state a claim yourself (category/assertion/
value); MEVA checks it against real recorded data and returns
SUPPORTED/CONTRADICTED/UNSUPPORTED/UNVERIFIABLE with full provenance. Full
details, including how the four modes differ: docs/playground.md.
See See MEVA in action near the top of this README for screenshots of the hosted sandbox (Guided Mode, a verification result, and Advanced Mode).
Documentation
Doc | Covers |
What MEVA is and is not โ read this first | |
Synthetic patient data provenance | |
The MCP tool layer | |
How MEVA talks to local Ollama models | |
The deterministic verifier | |
What reproducibility settings do/don't guarantee | |
The benchmark engine | |
Dataset construction and validation | |
Multi-model comparison methodology | |
Why END_TO_END and DECOUPLED both exist | |
The claim-extraction schema contract | |
The observation-category sanity audit | |
Full benchmark v0.3 report (historical) | |
Public fixture generation provenance | |
Why/how the former patient set was replaced | |
The public deterministic-verifier playground (CLI) |
Scope and safety
MEVA uses only synthetic (Synthea-generated) patient data โ no real
patient data is included or should ever be contributed. It performs no
diagnosis and no treatment recommendation, and is not clinically
validated. Its metrics (Evidence Grounding Score, Verifiable Claim
Coverage, etc.) are engineering/research benchmark metrics โ they
measure whether a model's claims match retrieved evidence, not medical
correctness, diagnostic accuracy, or patient safety. All AI inference is
local-only, through Ollama โ MEVA never calls a paid or cloud AI API.
Full statement: docs/safety-and-scope.md.
Contributing
Pick an issue (or propose one)
Fork the repository
Create a branch
Make your change
Run
pytestOpen a pull request
Full setup, testing details, and how to add FHIR support, benchmark cases
(synthetic data only), verifier tests, or model adapters:
CONTRIBUTING.md. Please also read CODE_OF_CONDUCT.md.
Looking for a first contribution?
Check the open GitHub issues: https://github.com/Tanz2024/meva-health-ai/issues
A good place to start is an issue labeled good first issue or help wanted.
See docs/contributor-issues.md for additional proposed contribution areas
and CONTRIBUTING.md for setup and submission guidelines.
License
MEVA's source code and locally-generated synthetic data are licensed under
the Apache License 2.0. Third-party dependencies and models have
their own licenses โ see THIRD_PARTY_NOTICES.md.
An earlier public-redistribution licensing question (a former patient set
copied from a repository with no declared license) was resolved in Stage
8A.1 by replacing that data with locally-generated Apache-2.0 fixtures โ
see docs/historical-sample-data-provenance.md for the full history.
Citation
See CITATION.cff for citation metadata for MEVA v0.1.0.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Servers
- Alicense-qualityCmaintenanceA clean-room SHARP-on-MCP compliant FHIR R4 MCP server that enables AI agents to interact with any FHIR R4 endpoint using SHARP context headers, without server-side OAuth. It provides clinical tools, lab results, imaging, and interactive MCP-UI dashboards.MIT
- Alicense-qualityDmaintenanceClinical decision-support MCP server that lets AI agents reason over live FHIR patient data for medication review, appointment scheduling, and care gap identification.7,813MIT

Agent-Townofficial
Alicense-qualityCmaintenanceA neutral verification court for AI tools that ranks MCP servers by executing them against ground truth and recording results. Enables agents to consult execution records, contribute verdicts, and challenge claims.Apache 2.0- AlicenseAqualityBmaintenanceA Claude-compatible MCP server that exposes health-domain tools over 100% synthetic data, built with security and compliance in mind.4MIT
Related MCP Connectors
Hosted MCP endpoint with realistic fake data for prototyping agents. 12 tools, no setup.
Deterministic fact verification for AI agents โ checksums & curated data, not guesses.
Read-only MCP over an agentic SLR workspace with per-claim citation verification
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/Tanz2024/meva-health-ai'
If you have feedback or need assistance with the MCP directory API, please join our Discord server