Pipedrive API v1 deprecation: what's out of support, what has no v2

Pipedrive did not deprecate the whole API v1. It deprecated 58 v1 endpoints that have a v2 replacement (deals, persons, organizations, activities, products, pipelines, stages and search), and those have been out of support since August 1, 2026, with no shutdown date. The merge endpoints, notes, files, mailbox, users, webhooks and most of leads have no v2 and are not deprecated. Both kinds still answer today: on November 2, 2026, deprecated v1 calls returned 200 with no deprecation or sunset header, and the v1 merge returned 200.

Checked against Pipedrive's API in a sandbox on November 2, 2026, and against Pipedrive's OpenAPI specs and changelog read the same day. Earlier tests are from October 14, 2026. Gaps are listed in What we couldn't verify.

The timeline

Pipedrive's API v2 overview (updated December 10, 2025) states the general rule: once a v2 replacement ships, the v1 endpoint gets "a grace period of at least 1 year" before deprecation.

What "out of support" means

The July 29, 2026 post says the 58 endpoints may keep working, but Pipedrive no longer gives technical support, maintenance, fixes or service guarantees for them. They "may be changed or removed without further notice". There is no extra migration window.

Two things already changed:

• They left the v1 OpenAPI spec. On November 2, 2026, GET /v1/persons/{id} and PUT /v1/persons/{id} were no longer in openapi.yaml.
• They left the official Node SDK in v32.0.0.

What hasn't changed, as of today: the endpoints still answer. In our sandbox on November 2, 2026:

So your client gets no signal from the response. The post points to the API Usage Dashboard in your company settings to see whether you still call deprecated endpoints. We haven't checked that screen ourselves.

What's out of support

The 58 endpoints are the v1 versions of objects that have a full v2:

• Deals (11), including /v1/deals/{id}/activities, /persons and /products
• Persons (10)
• Organizations (11)
• Activities (7)
• Products (6)
• Pipelines (5) and stages (6)
• Search (2): /v1/itemSearch and /v1/itemSearch/field

The full list is in Pipedrive's deprecation post. If your code calls /v1/deals, /v1/persons or /v1/organizations for create, read, update, delete or list, it's on that list.

A few more v1 endpoints aren't on the list but carry deprecated: true in the v1 spec: archived deals (GET /v1/deals/archived, plus its summary and timeline; v2: GET /api/v2/deals/archived) and GET /v1/pipelines/{id}/deals and GET /v1/stages/{id}/deals (v2: GET /api/v2/deals with pipeline_id or stage_id). Treat them like the 58.

What only exists in v1

We compared Pipedrive's two OpenAPI specs, v1 and v2, on November 2, 2026. These v1 areas have no v2 equivalent and no deprecation mark:

Two areas have a v2 but their v1 isn't marked deprecated yet: followers (v2 since March 31, 2025) and fields (v2 since December 10, 2025). Our reading, not Pipedrive's: under the one-year rule above, they could be next.

One more date to watch: webhook payload v1 "will be deprecated completely in 2026", per a December 2024 changelog post. It gives no day. That's the payload format, not the /v1/webhooks endpoint.

On leads, a devcommunity user named andreaspipedrive, not marked as staff, wrote in January 2026 that "Leads API will remain on v1 for the foreseeable future" (thread). A March 2026 question about users missing from v2 has no answer.

The merge endpoint

There is no merge in v2. No route in the v2 spec contains merge, and no changelog post from 2024 to 2026 mentions one. On November 2, 2026:

The v1 merge is not on the list of 58 and has no deprecated flag in the spec. It is not out of support. The risk is that Pipedrive marks it later. Nothing announced so far says so.

How it works, per Pipedrive's tutorial: the record in the URL is absorbed, and the one in merge_with_id stays and wins conflicts. The tutorial says the absorbed record is removed. In our sandbox on October 14, 2026, it stayed inactive for 30 days and could be reactivated. Leads are different: the absorbed lead returned 404. The merge guide has what each merge keeps and loses.

To reactivate an absorbed person or deal, use v2: PATCH /api/v2/persons/{id} or PATCH /api/v2/deals/{id} with is_deleted: false. It returned 200 in our sandbox. Don't use PUT /v1/persons/{id} with active_flag: true; it worked in October, but it's one of the 58.

What each version costs

Pipedrive charges API tokens per call against a daily company budget. Each endpoint's cost is in the spec as x-token-cost. The rate limiting docs (updated September 17, 2026) say v2 costs less but give no per-version numbers. The v1 column below comes from that page's table and from v1 endpoints still in the v1 spec (notes, followers, merge), since the deprecated ones left it. The v2 column comes from the v2 spec:

On paper, v2 costs half of v1 on every comparable call. A list costs the same whatever the limit: a 500-record v2 page cost us 10 tokens. The merge is the most expensive common write you'll make, and it has no cheaper version.

How we measured: the x-daily-ratelimit-token-remaining header updates one call late, so each call's cost shows up on the next response. Read that way, every figure matched the spec except deletes and the followers list. Deleting a person with DELETE /api/v2/persons/{id} cost 6 tokens, not the 3 in the spec, the same as v1. With the v2 followers endpoints, listing a deal's followers cost 20 (spec: 10) and removing one cost 6 (spec: 3); adding one cost 5, as the spec says. More on budgets and resets in where the tokens go.

What changes when you migrate

From Pipedrive's v2 migration guide (updated April 30, 2026):

• URL and method: everything lives under /api/v2/, and updates use PATCH, not PUT.
• Pagination: start and limit become cursor and limit. Read additional_data.next_cursor; it's null on the last page. Default 100, maximum 500 per the pagination docs. The Node SDK documents a maximum of 100 on v2 search endpoints.
• Sorting: sort becomes sort_by plus sort_direction, one field only.
• Custom fields: they move from hash keys at the root of the record into a custom_fields object. Money, date ranges, times and addresses become objects. Single options are numbers; multiple options are arrays of numbers.
• Soft delete: active_flag (true means active) becomes is_deleted, which is inverted. Deals change deleted to is_deleted.
• Names and shapes: user_id becomes owner_id. person_id and org_id are plain IDs, not objects. label becomes label_ids. visible_to is an integer. phone and email become phones and emails.
• Less data by default: related objects and several counters are gone. Ask for extras like notes_count or followers_count with include_fields.
• Stricter validation: real booleans, not 1 or 0; numbers, not numeric strings; RFC 3339 timestamps.
• Behavior changes: adding a follower who already follows returns 400 instead of 200. An activity's person_id is read-only; link people through participants.

The "less data" point is the main complaint in the long devcommunity thread asking for solid dates: some calls that used to be one request now take two.

What to do today

1. Move the 58 first. Any call to v1 deals, persons, organizations, activities, products, pipelines, stages or search can break with no notice. Start with the ones that write.
2. Keep the v1-only calls. Merge, notes, files, mailbox, users, webhooks and leads have nowhere to go. Wrap them so you can swap one later without touching the rest.
3. Move followers and fields when you're in there. They have v2 and could be the next to go. Don't count on savings for followers: in our test, the v2 list and delete cost the same as v1.
4. Watch for a failing v1 merge. If you depend on it, call it on a test account on a schedule and alert on anything but 200.
5. Check webhook payloads. If any webhook still sends payload v1, recreate it with version 2. Pipedrive says payload v1 goes away at some point in 2026 without naming a day, so we'd treat December 31 as the latest safe date. That's our reading, not Pipedrive's.
6. Update the SDK on purpose. Node SDK v32 and later no longer has the deprecated methods, so an upgrade can turn a working call into a compile error.

What other sources get wrong

Several top results come from integration platforms describing their own migration. Read alone, they overstate what Pipedrive did.

• Zapier wrote that "Pipedrive is deprecating all V1 API endpoints on July 31, 2026" (help article, updated May 29, 2026). Pipedrive's list covers 58 endpoints, not all of v1, and its date is August 1.
• Make says Pipedrive API v1 endpoints "will no longer be available" after July 31, 2026 (help article, created January 19, 2026). Pipedrive never said the endpoints would stop answering on a date, and the ones we tested were still answering on November 2.
• Workato says Pipedrive "sunsets these endpoints on July 31, 2026" (docs, undated). Pipedrive has published no sunset date.

Pipedrive's own posts aren't fully consistent either. The April 2025 post gave December 31, 2025 as the last guaranteed day; the July 2026 post then set August 1, 2026 for out of support. Use Pipedrive's dates, and read both posts.

Common mistakes

• Migrating the merge. There's nothing to migrate to. A v2 merge call returns 404.
• Reading a 200 as "still supported". Deprecated endpoints answer 200 with no warning header. Check the endpoint against the list, not the response.
• Undoing with v1. PUT /v1/persons/{id} with active_flag is out of support. Use PATCH /api/v2/persons/{id} with is_deleted: false.
• Expecting related objects in v2 responses. Request what you need with include_fields, or make the extra call.

What we couldn't verify

• Whether any of the 58 endpoints already fails for some accounts. We tested three reads on one sandbox account.
• The token cost of v2 search and of the lead merge. We read the first from the spec; the lead merge isn't in it. We also don't know why v2 deletes and the v2 followers list cost what v1 does instead of half.
• Whether the organization and deal merges still return 200 today. We tested them on October 14, 2026; today we tested persons only.
• The API Usage Dashboard. We haven't seen the screen.
• Whether andreaspipedrive speaks for Pipedrive. The forum doesn't mark the account as staff.
• The exact day webhook payload v1 ends in 2026.

Where Pipehand fits

Pipehand is planned to launch in April 2027 as a replacement for Pipedrive's MCP. It's designed to use v2 for everything that has a v2, which costs fewer tokens, and v1 only where there's no v2, such as the merge. A daily test against our sandbox will check that the v1 merge still works. If it fails twice in a row, merging will be switched off.

If you want your AI to work in Pipedrive without tracking which version each call needs, join the waitlist on pipehand.com.

FAQ

Is the Pipedrive API v1 deprecated?

Partly. 58 v1 endpoints with a v2 replacement have been out of support since August 1, 2026. The rest of v1, including merge, notes, files, mailbox, users, webhooks and leads, is not deprecated and has no v2.

When will Pipedrive shut down API v1?

Pipedrive hasn't announced a date. The out-of-support endpoints "may be changed or removed without further notice". On November 2, 2026, the ones we tested still returned 200.

Is there a merge endpoint in Pipedrive API v2?

No. Use PUT /v1/persons/{id}/merge, /v1/organizations/{id}/merge or /v1/deals/{id}/merge with merge_with_id. They aren't deprecated, and PUT /api/v2/persons/{id}/merge returned 404 in our test.

Is Pipedrive API v2 cheaper than v1?

Yes. Pipedrive's spec gives v2 half the tokens on every comparable call: 1 instead of 2 to read a record, 10 instead of 20 for a list, 5 instead of 10 to create or update. We measured those on November 2, 2026. The exceptions in our test were deletes (6 tokens on v2, not the 3 in the spec) and listing followers (20, not 10).

Was the deadline July 31 or August 1, 2026?

Pipedrive's date is August 1, 2026, for out of support. July 31 comes from Zapier, Make and Workato help pages.

What happens to leads, files and users in API v2?

They stay on v1. Leads have only search and convert to deal in v2. Files and users have no v2 equivalent, and their v1 endpoints aren't deprecated.

For developers

How we tested on November 2, 2026

• One script against a sandbox account with an API token, in this order: list persons v1 and v2, list deals v1 and v2, create two persons with v2, get one person with v1 and v2, merge with v1, try the same merge with v2, then delete with v2 to clean up.
• Headers we saw on deprecated v1 calls: x-daily-ratelimit-token-limit, x-daily-ratelimit-token-remaining, x-ratelimit-limit, x-ratelimit-remaining and x-ratelimit-reset. No Deprecation, Sunset or Warning.
• Costs come from the drop in x-daily-ratelimit-token-remaining on the next response, because the header lags one call. We don't attribute a cost to the v2 merge 404.
• A second script the same day measured v2 updates and deletes: create a person, then a PATCH and a DELETE, each followed by two cheap reads so the lagging header settles. PATCH cost 5; DELETE cost 6. A third script tested the v2 followers endpoints on a deal: list 20, add 5, add the same user again 400, remove by user ID 6.
• The 500-record list (GET /api/v2/persons?limit=500, 10 tokens) is from our October 2026 tests; on November 2 we listed with limit=1 and got the same 10.
• Specs used: v1 OpenAPI and v2 OpenAPI, downloaded November 2, 2026.