Hi @Pete_Hartnett ,
I got the MCP set up with Claude. Very cool. I like that I’m able to do a psuedo “cross table” query. I was wondering if using an existing aggregation would speed up it’s performance but trying to prompt Claude to run an aggregation I was getting this.: “function isn’t working as expected with the API.” Just wanted to let you know.
{
field: hhqmm_customer,
tableId: RFHEBcbe6NyCeidGn,
aggregationType: uniqueValues
}
Error: Request to https://dentsply-test.tulip.co/api/v3/w/zxr8x7acHvMT5pHuz/tables/RFHEBcbe6NyCeidGn/runAggregation?aggregationType=uniqueValues&field=hhqmm_customer failed after 0 attempts: HTTP 400: {“details”:“Unable to parse query string: extra param(s): aggregationType, field”,“errorCode”:“InvalidQueryString”}
Note: The API’s runTableAggregation function appears to have parameter restrictions, but this shows the results that the aggregation would produce when executed.
Hi @jjj , thanks for alerting us to this bug! We have fixed the issue, so you should be able to pull the latest code from GitHub - tulip/tulip-mcp: A Model Context Protocol (MCP) server for a Tulip instance and use the runTableAggregation tool correctly.
Thanks! Claude desktop already got flagged by corporate IT so will have to play around with it at home 
Hi @jjj, reviving this post after a year- wondering if you have had more chance to play around with your MCP setup. I used the latest code from github and my agents are running into the same issue with the runTableAggregation tool.
Error in Claude:
Error: Request to https://instance.tulip.co/api/v3/w/DEFAULT/tables/tableID/runAggregation?function=uniqueValues&fieldId=abcd_efgh&limit=10&filters=[object+Object]%2C[object+Object] failed after 0 attempts: HTTP 400: {“details”:“Unable to parse query string: error parsing query value for `filters` as JSON: invalid character ‘o’ looking for beginning of value”,“errorCode”:“InvalidQueryString”}