CRM API v3 Developer Preview is Now Available

[mwelch]

HubSpot API consistency and completeness has been a major point in feedback from our developer community for some time now. You spoke, and we listened! We are excited to announce the launch of a refreshed CRM API and documentation, available today in developer preview.

The CRM API v3 is the first milestone in a larger concerted effort to level up HubSpot’s APIs. These new APIs are more consistent than the current generation and are on their way to becoming more complete.
CRM Search, Imports, Owners, Pipelines, Properties, Associations, Objects (contacts, companies, deals, etc.)

Here’s what's new:
Search, Filter and Sort:
This brand-new endpoint allows developers to filter, sort, and search across all CRM objects (contacts, companies, deals, etc). These new capabilities will allow for more effective data retrieval across the CRM.

You might use this API for:

• Getting a list of contacts for a specific account
• Retrieving a list of all open deals
• Searching for contacts by custom properties

CRM Imports Endpoint:
This brand-new endpoint allows developers to import CSV files for creating or updating objects in their HubSpot accounts. These objects include contacts, companies, deals, tickets, products, and notes. The Imports API allows your integration to use the same powerful features that are supported by the in app import tools in HubSpot.

The Imports API allows you to:

• Import files directly
• Retrieve information on any past imports
• Cancel any incomplete imports

Learn once, apply everywhere:
All CRM objects now have more in common. Defining properties, associating objects to one another, assigning ownership and defining pipelines (where applicable) are all the same regardless of the objects you’re working with.

• CRM Object Properties
• CRM Associations
• CRM Owners
• CRM Pipelines

What this means for developers:

• Start testing out these endpoints. Developer preview is the perfect time to let us know what you think. We’ll be collecting and iterating quickly on your feedback, which could mean breaking changes. It’s not recommended to use these endpoints in a production environment yet, but we’ll let you know when they’re ready.
• Continue to use the existing stable CRM API, which will remain available for production use as we build out our platform with improved consistency and capability.

FAQ
CRM Search:

• Search functionality is limited to one request per second
• "engagements", including "tasks", "calls", "emails", and "notes" aren't supported in Search 

Coming Soon

• We are missing the Engagements and Contact Lists endpoints - use the existing ones for now.
• Secondary Identifiers - we’ll be adding this shortly—hang tight.

Please provide your feedback about these new APIs in this thread.

[mchanner]

Love the new API, how much longer till Engagements get added?

[Voyager2]

Is there a way to get deleted deals with the new API? It looks like it is not possible for the current version, one can only get deals with isDeleted: false

[Seth_Israel]

We were told this upgrade would help our customer association in Chat:

https://community.hubspot.com/t5/APIs-Integrations/How-to-identify-a-contact-in-the-chat/td-p/228672

Although since we don't use the API and instead use Segment to connect HubSpot, would we even see these upgrades on our end?

It looks like the last Segment HubSpot API upgrade was 16 months ago?
https://github.com/segmentio/analytics.js-integrations/tree/master/integrations/hubspot

[fndicu]

Hi Seth, 

Thanks for reaching out. When this functionality is in stable General Release (approx. end of q1) it will be ready for use in production environments. I recommend providing feedback to segment's development team. 

[allantl]

Very excited with the v3 API. The API seems more standardized.

I tried the Batch create contacts API /crm/v3/objects/contacts/batch/create.
I found a bug on the implementation.

Creating a contact with an email that already exists always returns an "id": "0". The same thing happens with single creation. If I remember correctly, it will throw an error "Email already exists" with v1?

Also, would it be possible to increate the Batch operation limits? 10 seems too little. What about increasing it to something like 25 or 50?

Another question that I have. According to the docs, this is a possible response.

{
  "results": [
    {
      "createdAt": "2019-10-30T03:30:17.883Z",
      "archived": false,
      "id": "512"
    }
  ],
  "status": "PENDING",
  "requestedAt": "2020-02-17T05:11:56.589Z",
  "startedAt": "2020-02-17T05:11:56.589Z",
  "completedAt": "2020-02-17T05:11:56.589Z"
}

How do we poll the status to check that it is COMPLETED? Do we need to rely on the webhook to get events (This would be so troublesome....)?

[fndicu]

Creating a contact with an email that already exists always returns an “id”: “0". The same thing happens with single creation. If I remember correctly, it will throw an error “Email already exists” with v1?

I've passed along this feedback to the team and it's something we’ve planned to address.

Also, would it be possible to increate the Batch operation limits? 10 seems too little. What about increasing it to something like 25 or 50?

For the time being, the limit will remain at 10, for now, to make sure the system is responsive and may increase it in the future. We will make a post on the changelog if the change occurs. 

How do we poll the status to check that it is COMPLETED

This was an unfortunate example in our API documentation reserved for future async patterns. All status responses in APIs available today should always read “COMPLETED”.

[allantl]

Hi fndicu,

Thanks for the response. I have another question for the batch creation API  /crm/v3/objects/contacts/batch/create

Based on the response json from the documentation, this means that some result can return an error? Or do you fail the whole batch if there's a validation error like the previous batch api in v1?

If some request can fail, how do I know which one of my request payload results in an error. Since there's no way for me to cross check in the `errors` object below. As a comparison, from Salesforce API .
What about returning the response based on the request payload order. Then, the success/error reason should be inside each object.

{
  "results": [
    {
      "createdAt": "2019-10-30T03:30:17.883Z",
      "archived": false,
      "id": "512",
      "properties": {
        "property_number": "17",
        "property_dropdown": "choice_b",
        "property_radio": "option_1",
        "property_string": "value",
        "property_multiple_checkboxes": "chocolate;strawberry",
        "property_checkbox": "false",
        "property_date": "1572480000000"
      },
      "updatedAt": "2019-12-07T16:50:06.678Z"
    }
  ],
  "numErrors": 0,
  "errors": [
    {
      "message": "Invalid input (details will vary based on the error)",
      "correlationId": "aeb5f871-7f07-4993-9211-075dc63e7cbf",
      "category": "VALIDATION_ERROR",
      "links": {
        "knowledge-base": "https://www.hubspot.com/products/service/knowledge-base"
      }
    }
  ],
  "status": "PENDING",
  "requestedAt": "2020-03-04T05:04:22.596Z",
  "startedAt": "2020-03-04T05:04:22.596Z",
  "completedAt": "2020-03-04T05:04:22.596Z"
}

[fndicu]

That is correct. The entire batch should run and we return an error for each of the specific objects that failed. 

[riwern]

Under the search endpoint there are some listed limitations. Are they planned limitations or just current limitations while the v3 API is under development ("unfinished")?

Archived CRM objects won’t appear in any search results.

In particular, I'm looking for a way to view contacts deleted ("archived=true") without pulling all contacts.

Secondary question: Will restore contacts be exposed within the API?

[fndicu]

Thanks for reaching out!

1) You can get a list of all archived contacts with GET /crm/v3/objects/contacts?archived=true, but archived contacts won’t appear as results when POST /crm/v3/objects/contacts/search is invoked.

 So you can list archived (deleted) contacts, but you can’t search for them (among other things, like update).

2) Yes, we plan to add this functionality as an update in the future!

[Ferseb]

Hi people,

I am new to the Hubspot API and I'm currently having a look at the V3 endpoints using the (rather good) Ruby gem. I can't seem to find anything related to the subscription of users to emails just like there was in v1 : https://api.hubapi.com/email/public/v1/subscriptions/MY_EMAIL?hapikey=KEY

Is it implemented yet?

Thanks for your help

[fndicu]

Hi Ferb, 

Glad to hear you are enjoying the experience so far! At this time we don't have that functionality implemented yet. Can you provide some more detail on your use case? Feel free to DM the details. 

[rossfairbairn]

Hi, firstly I just wanted to saying I'm generally loving the new API - thanks!

Something that is a bit of an issue, and please correct me if I'm wrong here but it doesn't seem possible to return associations when filtering tickets.

My user case: I need a list of tickets created in the last 7 days and for each ticket I need to know the company it's associated with.

Previously I had these two problems:

1. I had to get every ticket ever from the API and then filter down in my own script
2. I then had to get all the company -> ticket associations

This meant a lot of requests!

It seems you've solved both of these problems separately but not together. I can now return associations with tickets when getting all tickets but I can't return associations when using the new filter/search functionality.

[fndicu]

That is fantastic to hear! At this time that is correct, we can't return an object's associations in the same call. That being said, you can use the batch associations endpoint to return associations for a list of objects. 

I have passed along your feedback and will get it on the team's radar. 

Thanks!

[lukemelia]

In trying out the v3 contact search API, I am able to match a user by "email", but not by "id". e.g.

curl https://api.hubapi.com/crm/v3/objects/contacts/search\?hapikey\=APIKEY \
    --request POST \
    --header "Content-Type: application/json" \
    --data '{
      "filterGroups":[
        {
          "filters": [
            {
              "propertyName": "email",
               "operator": "EQ",
               "value": "luke@lukemelia.com"
            }
          ]
        }
      ],
      "properties": ["email", "firstname", "id"]
    }'

{
  "total": 1,
  "results": [
    {
      "id": "790301",
      "properties": {
        "email": "luke@lukemelia.com",
        "firstname": "Luke",
        "id": "d15231c3-7580-43c5-9ee7-5a7dc7a37707"
      },
      "createdAt": "2019-02-13T18:28:48.982Z",
      "updatedAt": "2020-02-14T01:03:19.448Z",
      "archived": false
    }
  ]
}

curl https://api.hubapi.com/crm/v3/objects/contacts/search\?hapikey\=APIKEY \
--request POST \
--header "Content-Type: application/json" \
--data '{
"filterGroups":[
{
"filters": [
{
"propertyName": "id",
"operator": "EQ",
"value": "d15231c3-7580-43c5-9ee7-5a7dc7a37707"
}
]
}
],
"properties": ["email", "firstname", "id"]
}'

{
"status": "error",
"message": "There was a problem with the request.",
"correlationId": "7533e820-6dbe-43a0-8155-d03c9783c05b",
"requestId": "2f034108-5bba-4c7e-851f-ebb8f4f4cb25"
}

Is this a bug or am I missing something?

[fndicu]

Hi Luke, I'll shoot you a dm to dig in a little further. The only strange thing I notice here is that it seems like we have two properties with the name "id" and two different formats. 

[nadendav]

The v3 Contacts Batch Endpoint does not appear to be performing an Upsert as I believe is described for the v1. I understand this to be related to de-duping based on existence of an E-mail address.

[nadendav]

Thank you, Thank you, Thank you for the v3 Contacts API. I had been banging my head against a wall trying to get my iPAAS solution to generate the v1 JSON mesasge body with no luck. It looks like I got the v3 to work yesterday! More testing today... 

I am also interested in the new v3 Imports Endpoint but soi far I haven't been able to put 2 and 2 together. I think I've studied the content at the links provided already and I can't find a specification for the importRequest and files parameters. The link at "See this page for more on the required format." for the importRequest does not go anywhere, it simply returns the same page. Thanks 

[fndicu]

We are glad to hear you are enjoying the new API! 

Thank you for the feedback on Imports. We are currently working on a fix that will provide the sample import post body in-line with the overview documentation. 

[nadendav]

Awesome! Can't wait...