
Query the AACT (ClinicalTrials.gov) database directly from Claude. Explore 70+ tables of clinical
Query the AACT (ClinicalTrials.gov) database directly from Claude. Explore 70+ tables of clinical trial data — studies, interventions, outcomes, sponsors, facilities — using read-only SQL with buffered pagination.
·
Query the AACT (ClinicalTrials.gov) database directly from Claude. Explore 70+ tables of clinical trial data — studies, interventions, outcomes, sponsors, facilities — using read-only SQL with buffered pagination.
The ClinicalTrials.gov API returns one JSON record per trial — useful for quick lookups, but awkward for analytics. Want the average duration of Phase 2 NSCLC trials from 2020-2025? With the API you'd filter trials, extract dates from each JSON record, then compute durations client-side. With AACT, that's a single SQL query.
A structured PostgreSQL database makes it far easier to aggregate, combine, and summarize clinical trial data in any way you need. And for AI-assisted analysis, SQL is a standard that LLMs handle extremely well — fewer mistakes, less context to manage, better performance, and lower cost compared to parsing bespoke API responses.
Note: This is an independent, third-party integration. It is not affiliated with or endorsed by the Clinical Trials Transformation Initiative (CTTI) or Duke University. However, we released a case study with CTTI on integrating their database with Claude - see AACT case study.
| Tool | Description |
|---|---|
database_info | Confirm database connection, server time, and data currency |
list_tables | Discover all available tables with approximate row counts |
describe_table | Inspect column names, types, distinct counts, and sample values |
get_column_values | Get distinct values for a column with counts — essential before filtering |
search_columns | Find columns by keyword across all tables (e.g. masking -> designs.masking) |
read_query | Execute a SELECT, CTE, or EXPLAIN query with buffered results and preview |
fetch_rows | Page through buffered query results without re-querying |
All tables join on nct_id.
Every read_query statement is parsed before it runs. Only one read-only SELECT, WITH, or EXPLAIN statement is accepted. It runs in a read-only database session and is cancelled after 2 minutes.
Download the latest .mcpb file from Releases and open it in Claude Desktop. You'll be prompted for your AACT credentials.
Add to your claude_desktop_config.json (~/Library/Application Support/Claude/claude_desktop_config.json on macOS, %APPDATA%\Claude\claude_desktop_config.json on Windows):
{
"mcpServers": {
"aact": {
"command": "uvx",
"args": ["mcp-server-aact"],
"env": {
"DB_USER": "your_username",
"DB_PASSWORD": "your_password"
}
}
}
}{
"mcpServers": {
"aact": {
"command": "docker",
"args": [
"run", "--rm", "-i",
"--env", "DB_USER=your_username",
"--env", "DB_PASSWORD=your_password",
"navisbio/mcp-server-aact:latest"
]
}
}
}git clone https://github.com/navisbio/AACT_clinicaltrials_MCP.git
cd AACT_clinicaltrials_MCP
uv sync{
"mcpServers": {
"aact": {
"command": "uv",
"args": ["run", "--directory", "/path/to/AACT_clinicaltrials_MCP", "mcp-server-aact"],
"env": {
"DB_USER": "your_username",
"DB_PASSWORD": "your_password"
}
}
}
}"Who are the top 10 sponsors of Phase 3 Alzheimer's disease trials? Break down by trial status."
The server will discover relevant tables, check enum values for phase and status, then build a query joining studies, conditions, and sponsors.
"Find all actively recruiting Phase 2 and Phase 3 trials for pembrolizumab in non-small cell lung cancer. Show NCT ID, title, enrollment, and lead sponsor."
Uses get_column_values to confirm phase format (PHASE2, PHASE3), then queries across studies, browse_interventions, and conditions.
"What are the most common primary outcome measures in completed Phase 3 type 2 diabetes trials?"
Joins studies with outcomes to analyze endpoint patterns, grouped by outcome measure type.
"How many clinical trial sites does a typical rare disease trial have? Show the top countries by site count."
Queries the facilities table joined with conditions to map trial geography.
This server is read-only and does not collect or store any personal data. See PRIVACY.md for details.
DB_USER and DB_PASSWORD are set correctly in your configspawn uvx ENOENT errorThe system cannot find uvx. Use the full path:
{
"mcpServers": {
"aact": {
"command": "/Users/username/.local/bin/uvx",
"args": ["mcp-server-aact"],
"env": {
"DB_USER": "your_username",
"DB_PASSWORD": "your_password"
}
}
}
}MIT
Pick your client and paste the snippet. Each one is the same server, written the way that client expects it.
claude mcp add aact -- uvx mcp-server-aact{
"mcpServers": {
"aact": {
"env": {
"DB_USER": "",
"DB_PASSWORD": ""
},
"args": [
"mcp-server-aact"
],
"command": "uvx"
}
}
}code --add-mcp '{"name":"aact","env":{"DB_USER":"","DB_PASSWORD":""},"args":["mcp-server-aact"],"command":"uvx"}'[mcp_servers.aact]
command = "uvx"
args = ["mcp-server-aact"]Runs locally on your device. Your client starts the server itself, so nothing has to be hosted.
Paste this prompt into your agent. It reads this page and does the setup for you.
Read https://aiagentslisting.com/mcp/aact-clinical-trials-mcp-server to learn what the "AACT Clinical Trials MCP Server" MCP server does and how to install it. Add it to my coding agent's MCP configuration as documented on that page, then confirm the server connects and list the tools it exposes.Agents can also browse this directory over MCP at https://aiagentslisting.com/api/mcp
This server runs locally, so we can't read its tool list over the web yet.
We could not run this server: it ships no npx package for our sandbox and no public endpoint. The four checks that need a live server are left out, and the grade tops out at B until one exists.
Maintain this server? Claiming proves you own the listing and earns the last 4 points.
Claim this listingPaste this line near the top of your repository's README. The badge always renders the current score.
[](https://aiagentslisting.com/mcp/aact-clinical-trials-mcp-server)Nothing comparable is listed yet.
v0.5.0v0.5.0 — Claude Desktop PluginRelease notesv0.4.0Release v0.4.0Release notesConnect to the AI Agents Listing MCP server at https://aiagentslisting.com/api/mcp using the HTTP transport and complete the sign-in when it asks. Then call the submit_review tool for the listing "aact-clinical-trials-mcp-server" with a rating from 1 to 5 and a short note on what worked and what did not.No reviews yet
Be the first to share your experience with this tool.
One email a week. New agents, MCP servers and skills, and what is actually getting traction.