twimco.Let’s talk ↗︎
Menu
Field notes / Operations

HubSpot turns off Pipelines API V1 on December 4. Do a dry run of that morning before it happens to you

On the morning of December 4 something in your HubSpot setup may stop saying which deal stage a lead belongs in. It won't announce itself. Walk through that day on paper now, while there's still time to find what's calling.

Editorial engraving of a sales operations lead moving blank index cards between columns on an office wall at morning while a colleague at a standing desk points to the same column.

HubSpot will retire its Pipelines API V1 on December 4, 2026, and calls to V1 endpoints will stop responding. If a form, sync, automation or script of yours reads or edits deal or ticket pipelines through V1, move it to the 2026-03 or newer version HubSpot recommends before then. Find the callers first: custom code, agency work, no-code tools with code steps, and vendor connectors you haven't looked at since they were set up.

What HubSpot announced

HubSpot posted a notice in its developer changelog on August 14: the Pipelines API V1 will be retired on December 4, 2026. After that date it won't be supported and calls to V1 endpoints will stop responding. If an integration of yours depends on V1, HubSpot says to migrate to its 2026-03 or newer version, and it recommends the 2026-09 version, which became generally available on September 9. The notice also says to audit your codebase for V1 calls.

Pipelines are how HubSpot organizes deals and tickets into stages, and code that reads or edits them is common: forms that fill a stage dropdown, syncs that create a deal in a named stage, reports that group by stage. The changelog entry doesn't list the affected endpoints or describe the error you'll see, so the first job is to find out whether you're affected at all.

Walk through the morning

It helps to read the day before it happens. This is an illustration, not a prediction, but it's the shape these things usually take.

Nothing is wrong at 9:00. The overnight sync ran, the forms loaded, the dashboards look the way they did yesterday. Around mid-morning, a web form that asks HubSpot for the list of stages to fill a dropdown shows an empty list. Someone from the website team notices and assumes a plugin glitch. At lunch a sales rep mentions that leads from the Thursday campaign aren't landing in the stage they usually do. In the afternoon, someone digging through an old project folder finds a script, written for a launch two years ago, that looks up a stage by its name every time it creates a deal.

The report that quietly counts deals by stage is the last to show it. It doesn't fail. It just comes up short, and it's a week before anyone asks why.

None of this needs bad luck. It needs one integration nobody remembered, built by someone who isn't there anymore, calling an endpoint that used to work.

Sequence of work

The morning of December 4, illustrated

  1. Before 9:00

    Nothing looks wrong. Overnight syncs may have run against V1 for the last time.

  2. Mid-morning

    A form that fills its stage dropdown from HubSpot shows an empty list or an error.

  3. Lunch

    A sales rep asks why new leads aren't landing in the right stage.

  4. Afternoon

    Someone finds a script from an old project that looks stages up by name.

  5. Next week

    A report that counts deals by stage comes up short and nobody knows why.

Illustrative scenario for a business with one old integration; not a record of any real incident.
You can't migrate a call you haven't found yet.

Where the old calls hide

Put the dry run aside and look in four places.

Your own code. Any custom app, serverless function or script that talks to HubSpot. Search for "pipelines" in the code and check which version path it uses. An old call usually has "v1" sitting right next to "pipelines" in the address, so a search for both together narrows things quickly. Don't forget configuration files and environment settings, where a base address is sometimes stored apart from the code that uses it.

Work an agency or freelancer did. The forms, the landing page integrations, the quick sync from a store. Ask for the repository, the access list and the last time anyone changed it. If nobody has one, that's a finding on its own.

No-code tools with a code step. Many automation tools let someone paste a small script into a step. Those steps are easy to forget because they sit inside a tool that looks like it's all drag-and-drop. Open each workflow that touches HubSpot and read every step that isn't a plain connector.

Connectors you didn't build. Marketplace apps and vendor integrations that sync deals into another system. Ask each vendor, in writing: "Do you use HubSpot's Pipelines API V1? When will you move off it?" Save the reply with a date.

The fix list for the week before

Do these in order, because the first one is the one people skip.

You can't migrate a call you haven't found yet. Make a list of everything that connects to HubSpot, with an owner next to each. Open each one and look for a Pipelines call. Where you find one, move it to the version HubSpot names, which means reading HubSpot's reference for the current shape of the request and response rather than only changing a number in the URL, since newer versions can differ in what they return.

Then test against something safe. If you have a HubSpot test account, copy your real pipeline and stage setup into it and run each integration there. Check the awkward cases: a stage that was renamed, a pipeline with an archived stage, a ticket pipeline as well as the deal one.

Finally, decide what "working" means before you flip anything: one deal created in the right stage, one dropdown filled, one report matching last week's numbers. Write it down so the person checking on December 4 knows what to look for.

Two practical points about timing. First, don't schedule the switch for the last day. Moving a week early leaves time to notice the report that comes up short before HubSpot makes the decision for you. Second, think about who is around. December 4 is a Friday in a busy season, and the week before it includes the Thanksgiving holiday for U.S. teams. If the person who knows the old sync is taking that week off, the realistic deadline is earlier than the one in the changelog.

Keep the old code in version history rather than deleting it on the day. If the new call returns stages in a different shape than you expected, being able to compare the two side by side saves an afternoon.

If you can't find the caller

Sometimes you're sure something is calling V1 and you can't tell what. HubSpot's own changelog says to audit your code, but it doesn't help with a tool you didn't write. In that case, work backward from symptoms: which forms, syncs and reports depend on pipeline stages, and which of them has a person who can say how it's built? It's also worth putting the Monday after December 4 in the calendar as a check, not just the day itself, since slow-failing reports show up later.

Most teams that find themselves here have the same underlying gap: nobody owns the map of what's connected to what. Building that map, and the automations on top of it, is the kind of work our business automation practice does. If a deadline like this one has made you nervous about your setup, talk to us.

Related work: Spin:Market

How the work moves

The week before

  1. Find

    List every place that talks to HubSpot's API, owner by owner.

  2. Read

    Open each one and look for a Pipelines call on the V1 path.

  3. Swap

    Move each to the 2026-03 or newer version HubSpot names.

  4. Test

    Run each against a test account with a copy of your real pipelines.

  5. Watch

    Check the first Monday after December 4 on purpose.

Decision framework for a one-week migration; not a HubSpot procedure.
twimco.

We build alongside the people doing the work.
Bring us the problem you’re thinking about.

Related work

Spin:Market ↗︎