Connect Claude to Pipedrive

Search deals by free text. `term` is required and must be at least two characters - Pipedrive has no "list every deal" search, so start from a name, a company or a keyword. Narrow with person_id or organization_id when you already know who the deal belongs to, and with status to separate open work from won and lost. Returns deal ids; pass one to get_deal for the whole record. Pages by cursor rather than by offset.

Fetch one deal in full, by the id search_deals returns. Ask for custom_fields when the account keeps anything important in them, and for include_option_labels when you want a dropdown's label rather than its numeric id. The returned stage_id is an internal id, which list_stages translates into a name.

Search people by name, email address or phone number. `term` is required and must be at least two characters. Set `fields` to "email" to look somebody up by address without matching their name as well. Returns person ids and the matching contact details; pass an id to get_person for the whole record.

Fetch one person in full - their name, every email address and phone number on the record, their employer, owner and custom fields. Takes the id search_persons returns.

Search organizations by name or address. `term` is required and must be at least two characters. Returns organization ids; use one as org_id on a deal or a person, or as organization_id when narrowing search_deals or search_persons. Unlike the deal and person searches this one takes no owner filter and no include_fields.

List activities - Pipedrive's calls, meetings, emails and tasks - newest change first. Narrow by deal_id, person_id, org_id or owner_id to see one thing's history, and by done to separate what is outstanding from what is finished. Note what v2 does not have: there is no start_date or end_date window any more, so bound a query with updated_since and updated_until, or sort by due_date and page.

List the account's pipelines, each with its id and name. A pipeline id is what create_deal and update_deal take, and it is also the filter list_stages needs. This is account configuration - it holds no deal and no person.

List pipeline stages with their ids, names and order. This is the translation table between what a person says and what the API takes: "Negotiation" is a name, and create_deal and update_deal only accept the numeric stage_id. Pass pipeline_id to see one pipeline's stages rather than every stage in the account. Account configuration - no deal and no person appears here.

List the account's own Pipedrive users, with their ids, names and email addresses. Call it to turn a colleague's name into the owner_id that create_deal, update_deal, create_activity and add_note take. These are the caller's own colleagues rather than anybody the account sells to, which is why the response stays in the audit log. It is the connection check as well, because it takes no parameters at all and reads no record. This is the second of the two operations still on API v1 - Pipedrive built no v2 user list.

Create a deal. Only `title` is required; everything else is optional and Pipedrive fills in the account's defaults. pipeline_id and stage_id are internal ids rather than the names shown in the UI - call list_pipelines and list_stages first and copy the ids from there. owner_id is the Pipedrive user the deal belongs to, from list_users. These are the v2 field names: it is org_id and owner_id, not the v1 organization_id and user_id.

Change fields on an existing deal - this is how a deal moves to another stage, gets re-priced, or is marked won or lost. Only the fields you send are touched. v2 uses PATCH here; v1's PUT is gone. stage_id must be an internal id from list_stages, not a stage name. This cannot delete the deal.

Create a person. Nothing is strictly required, but a record with no name is not useful. Note the v2 shape for contact details: `emails` and `phones` are lists of objects, each {"value": "[email protected]", "label": "work", "primary": true} - v1's single `email` and `phone` strings are gone. Search first; Pipedrive will happily create a second record for somebody who is already there.

Schedule an activity - a call, a meeting, a task - and attach it to a deal, a person or an organization. `type` is the account's own activity type key from its settings, e.g. "call", "meeting", "task"; "call" and "meeting" exist in every account. due_date is a plain date and due_time a 24-hour clock time, both in the account's time zone. This is internal scheduling only: Pipedrive's calendar attendee list, which mails an invitation to an outside address when calendar sync is switched on, is deliberately not offered here, so nothing this operation does is visible to anybody outside the account.

Log a note against a deal, a person or an organization. The note appears on that record's timeline for the account's own users and is not sent to anybody outside it, which is why this is a write. `content` is HTML, so line breaks want <br> rather than \n, and Pipedrive sanitises what it is given. Pipedrive requires at least one of deal_id, person_id, org_id or lead_id - a note attached to nothing is refused. This is one of the two operations in this connector still on API v1: Pipedrive never built a v2 for notes, and the v1 endpoint remains supported.