fetch
ChatGPTFetch a HubSpot record via MCP.
Business
HubSpot (Legacy)
Overview
Analyze CRM data and surface insights
Tools
oai_user_fetch
ChatGPTUses mfetch to perform document fetch for link following and citations.
oai_user_search
ChatGPTUses synced RAG index to perform semantic search against query. The index is refreshed until (now - 2h).
search
ChatGPTSearch HubSpot data using the MCP service. The connector expects the HubSpot Search DSL and will reject natural-language fragments. Every query string must: Begin with `_query_version_2_ op:<describe_object|find_objects|count_objects> object_type:<contacts|companies|deals|tickets>`. Continue with space-separated
key[:op]:value tokens only. Limit keys to `q`, `query`, confirmed HubSpot properties, `property_names`, `limit`, `offset`, `sort`, and `associated_<object_type>`. Limit operators to eq, neq, gt, gte, lt, lte, in, not_in, contains_token, not_contains_token, has_property, and not_has_property (for associated_* only use eq, in, not_in). Use only `q:"text"` or `query:"text"` for full-text search and never embed bare prose elsewhere in the DSL. Apply pagination or property_names tokens only when necessary and use sort:<property>:asc|desc for ordering. Unknown schemas should be handled by first issuing a _query_version_2_ op:describe_object object_type:<type> limit:10 request before filtering on properties. Examples: - _query_version_2_ op:describe_object object_type:contacts limit:10 - _query_version_2_ op:find_objects object_type:contacts q:"john doe" limit:10 sort:createdate:desc - _query_version_2_ op:find_objects object_type:deals dealstage:not_in:closedlost,closedwon closedate:gte:2025-01-01 sort:closedate:ascApp Stats
4
Tools
3
Prompts
May 11, 2026
First seen
ChatGPT
Platforms
Category
CRM RecordsWorks with
ChatGPT
Data refreshed daily