jump_get_document_content
ChatGPTReads the full text of one to ten documents by id, as Markdown extracted from the original file. max_content_length caps the text per document; null returns everything. Documents still being processed return placeholder text.
jump_get_meeting_content
ChatGPTReads one to ten meetings by id. detail_level chooses how much comes back: metadata (title, date, participants, link), summary (adds the generated notes and action items), or notes (the complete generated notes as Markdown, as shown in Jump). Full transcripts and recordings are not available through this connector; to find what was said, use jump_semantic_meeting_search. Participants come as up to ten names with a total count, the quickest way to confirm who attended a meeting found with another tool. Calendar events without a recording are reported as missing.
jump_list_documents
ChatGPTLists the user's documents by name, linked contact, document type, or date, with title, type, upload date, contact, and link. Defaults to the last 90 days. Does not search inside documents; use jump_semantic_document_search for that.
jump_list_jump_tasks
ChatGPTLists tasks tracked in Jump: follow-ups generated from meetings, emails, documents, and forms, plus tasks created by hand. Filter by status, due-date window, overdue only, keyword, source meeting, or assignee, for the current user or the whole team. Returns title, description, due date, status, source, assignee, and a link. Tasks that live in a connected CRM are separate records and are not included.
jump_list_meetings
ChatGPTLists the user's meetings and calendar events in a date range with title, start time, up to five attendee names, whether a transcript exists, and a link. Newest first unless order is asc. Defaults to the last 90 days; one date bound expands to a 90-day window. Filter with meeting_kind: past, upcoming, recorded (meetings with a transcript, any date), or all. Use jump_get_meeting_content to read the summary, notes, or action items of a listed meeting.
jump_list_participant_meetings
ChatGPTFinds meetings by the people in them. Give contact_ids to list past recorded meetings that include every one of those saved contacts, or give search with a name or email to match meeting participants and calendar attendees, including people who are not saved contacts. Use one or the other, not both. Returns each meeting's time, its id when it has one, and the person that matched. Does not search what was said; use jump_semantic_meeting_search for that, and jump_get_meeting_content to read a meeting found here.
jump_search_jump_contact
ChatGPTFinds contacts saved in Jump by email, phone, name, or free text. Searches individual people by default; set contact_type to include households. Returns contact details such as name, email, phone numbers, addresses, date of birth, client tier, and linked CRM records, up to 30 matches. Email is the most precise lookup, then phone, then name; use search only as a free-text fallback. Results are text matches, so an ambiguous query can return several candidates.
jump_search_meeting_participants
ChatGPTFinds people who took part in the user's meetings, by email, name, or free text, including people who are not saved as contacts. Returns one record per person with name, email, how many meetings they attended, and the most recent meeting date. Does not return meeting content. Defaults to the last 90 days when no dates are given; one date bound expands to a 90-day window.
jump_search_participant_meetings
ChatGPTFinds meetings by a person's name or email within a date range. Matches meeting participants and calendar attendees, including people who are not saved contacts and upcoming events that have no recording yet. Both time_bounds.start and time_bounds.end are required. Returns each meeting's time, its id when it was recorded, and the person that matched. Does not search what was said; use jump_semantic_meeting_search for that, and jump_get_meeting_content to read a meeting found here.
jump_semantic_document_search
ChatGPTSearches inside the content of the user's documents by meaning. Returns matching snippets with the document title, type, linked contact, and relevance score. Use it for questions about what a document says; use jump_list_documents to find documents by name, contact, type, or date.
jump_semantic_meeting_search
ChatGPTSearches what was said and written in recorded meetings: transcripts, notes, and titles. The best first choice for any question about a topic, a decision, or a detail from a conversation. Returns matching snippets with the meeting id, title, and relevance score, but no participant list; use jump_get_meeting_content to see who attended. Transcript matches come back as short excerpts, a few per meeting at most, so ask a specific question rather than trying to read a whole meeting. query_mode picks semantic, hybrid, or exact keyword matching; hybrid or semantic is the safer default, the keyword modes drop everything that does not match exactly. meeting_topic narrows by meeting title: meeting_topic_use "rank" boosts matching titles while keeping content matches, "filter" keeps only matching titles. prioritize_recent boosts recent meetings; time_bounds sets a hard date range. If an exact search returns nothing, retry with a ranked one.
jump_get_document_content
Claudejump_get_meeting_content
Claudejump_list_documents
Claudejump_list_meetings_beta
Claudejump_search_jump_contact
Claudejump_search_meeting_participants
Claudejump_search_meetings
Claudejump_semantic_document_search
Claudejump_semantic_meeting_search
Claude