list_tablesLABSCO SUMMARY
Give the team the data, not the database.
Read-only and admin-scoped is the whole design. End users get natural-language access to exactly the tables that were exposed to them, and no data is copied anywhere to make that work.
What it is
A hosted, multi-tenant MCP server. Your organisation connects its data sources once; users then query them from Claude, ChatGPT or any Streamable HTTP client. There is nothing to install or run — you connect to an authenticated URL.
What you get
- The tables you have been granted, listed
- A table's columns and types inspected
- Curated metadata read for a table
- Sample rows pulled before you write anything
- Queries run against the exposed tables, read-only
- Collection queries for document-shaped sources
Requirements
A DataAssist-IO account and at least one connected data source. The endpoint is https://app.dataassist.io/mcp over Streamable HTTP, with OAuth and Dynamic Client Registration, discoverable at /.well-known/oauth-protected-resource. It is also published as a ChatGPT app.
Setup effort
Follow the published steps — Create a DataAssist-IO account and organization at https://dataassist.io
USE CASES
01
Letting a colleague who does not write SQL answer their own question
They ask in plain language and it runs against only what they have been granted.
02
Finding out what you actually have access to
List the tables available to you before asking anything else.
03
Writing a query against an unfamiliar table
Inspect its columns and types, then look at a few sample rows.
04
Understanding what a column really means
Read the curated metadata instead of guessing from the column name.
05
Querying document-shaped sources alongside tables
Run a collection query over the same connection.
6 tools
Listed in project docs. Your client may show a different list.
describe_tableget_table_metadataget_sample_dataquery_dataquery_collectionBEFORE YOU INSTALL
What to know first
Tools you get
6 tools named in the project's documentation.
INSTALLATION
Connect your client
Server setup
Create a DataAssist-IO account and organization at https://dataassist.io
Connect a data source — upload a CSV or connect a database/warehouse from the dashboard
Expose tables to your team and add curated metadata (optional but improves answer quality)
In Claude (Desktop or web), add a custom connector pointing at https://app.dataassist.io/mcp — or open the published ChatGPT app; any Streamable-HTTP + OAuth client can point at the same endpoint
Complete the OAuth sign-in the client walks you through; your org's tables then appear as queryable tools
