list_collectionsLABSCO SUMMARY
Genomics portals, queryable in a sentence.
Instead of learning five portals you ask what collections exist, what a table's fields are, and how many variants match — then drop into SQL when the question gets specific.
What it is
A Node server wrapping the Omics AI Python library. It browses collections and tables across Omics AI Explorer networks, reads their schemas, and queries them with filters or with SQL.
What you get
- list_collections discovers what a network holds; list_tables browses inside a collection by collection_slug
- get_schema_fields returns a table's fields and their types before you write a query
- query_table queries with filters, pagination and order_by
- count_rows counts what matches without pulling the rows back
- sql_search runs Trino SQL directly for anything the structured path will not express
- Reaches HiFi Solves, Neuroscience AI, ASAP, Viral AI and Target ALS by name, or any network by URL
Requirements
Node and npm: clone the repository and run npm start. Public collections need no credentials — access_token is an optional parameter for the ones that do.
Setup effort
Build from source — clone the repository, then follow the project's own run instructions
USE CASES
01
Finding out what data a network actually holds
Collections and the tables inside them come back by name, which is the step that normally means reading a portal's documentation.
02
Reading a table's schema before writing a query
Field names and types come first, so the query that follows is written against what is there rather than what was assumed.
03
Pulling variants that match a condition
Filters, ordering and pagination on a table, so a first look at a dataset does not require a download.
04
Counting matches without downloading them
count_rows answers how many before you decide whether the extract is worth pulling.
05
Asking something the filters cannot express
sql_search takes Trino SQL directly, which covers joins and aggregates the structured path leaves out.
6 tools
Listed in project docs. Your client may show a different list.
list_tablesget_schema_fieldsquery_tablecount_rowssql_searchBEFORE YOU INSTALL
What to know first
Tools you get
6 tools named in the project's documentation.
Requires
node on PATH
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
Server setup
TERMINAL
node src/index.jsMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"omics-ai": {
"command": "node",
"args": [
"src/index.js"
],
"cwd": "<YOUR_PATH_TO_OMICS_AI_MCP>"
}
}
}
