Menu

Mattering Automations Capsule CRM integration

Connect Capsule CRM to Mattering Automations: people and organisations, opportunities and pipelines, projects and boards, tasks, tracks, history entries and the account configuration behind them, as workflow steps authorized by a single bearer token that any Capsule plan can generate.

The Mattering Automations Capsule CRM integration puts Capsule's API v2 on your automation canvas: create and update people and organisations, open and move opportunities through pipelines and milestones, run delivery work as projects across boards and stages, create and complete tasks, apply whole task sequences with tracks, write notes onto a record's history, and read or edit the custom fields, tags, activity types, teams, users and goals behind all of it. 131 nodes in total, 67 that read and 64 that act, covering 130 of the 132 operations Capsule documents. The two that are not built are both attachment file transfers, and the node count runs one ahead of the operation count because Capsule's single party create endpoint is split into a person node and an organisation node.

The credential is a single bearer token. Any Capsule user can generate a personal access token from My Preferences then API Authentication Tokens inside their account, and Capsule's plan table includes API Access on every plan including the free one, so there is no upgrade gate on connecting. There is no OAuth flow to run and no application to register for ordinary use. Two operations do need more than a personal access token can offer, and both say so on the node rather than failing mysteriously: updating a user's preferences needs the user_preference scope, and subscribing or unsubscribing a REST hook needs a registered application authenticated through the full OAuth flow plus an Administrator role.

Three things about Capsule are worth reading before you build anything, because each one produces a workflow that runs green and does nothing useful. A party is not a contact: Capsule stores people and organisations in one polymorphic collection with disjoint name fields, so a template wired to the wrong one silently blanks half the account. Milestones belong to pipelines and stages belong to boards, and they are two separate hierarchies rather than two names for one thing. And tags and custom field values are omitted from every party, opportunity and project unless a node's embed asks for them, so a workflow reading a custom field must request it. All three are explained below, and this guide ends with the complete node reference.

Before you start

  • A Capsule CRM account. Any plan will do: Capsule's plan comparison table lists API Access with a tick in the Free, Starter, Growth, Advanced and Ultimate columns, so a free account can generate a token and connect.
  • An awareness of what the free plan caps, because those limits produce plan errors from Capsule rather than authentication errors. A free account is capped at 250 contacts and 5 custom fields, and includes 1 sales pipeline, 0 project boards, 0 teams and 0 email templates, with no goals, no customizable sales milestones and no custom activity types. Project boards start on Starter and teams start on Growth.
  • A decision about which Capsule user holds the token. Every node runs as that user and inherits their access, and unattended automations keep running on it, so a user the business controls beats an individual's personal login.
  • The Administrator role on that user if your workflows will write account configuration. Creating or updating boards, stages, lost reasons and custom titles requires it, and so does managing REST hook subscriptions.
  • A registered Capsule OAuth application, only if you intend to use Capsule: Subscribe REST Hook or Capsule: Unsubscribe REST Hook. Capsule states that only registered applications authenticated through the full OAuth flow can manage subscriptions. Everything else in this integration works with a personal access token.
  • A Mattering workspace with access to the Automations app.

Generate a Capsule personal access token

Capsule offers two ways to obtain the bearer token its API expects, and this integration accepts either. For a one-off internal integration Capsule's own recommendation is a personal access token generated straight from your account, which avoids the OAuth flow entirely and carries no documented expiry. The alternative is an OAuth 2 access token, which Capsule issues for seven days and which this integration does not refresh for you, so the token you paste stops working after a week. Unless you have a specific reason to run OAuth, use a personal access token.

Sign in to Capsule as the user the automations should act as.
Open the profile menu, go to My Preferences, then open API Authentication Tokens.
Click the option to generate a new token, give it a name that says which system will use it, and generate it.
Copy the token now and store it somewhere safe. Treat it as a password: it authenticates as the whole user, it has no documented expiry, and anyone holding it has that user's access to the CRM.
Capsule names no plan requirement and no role requirement for generating a personal access token, so this step works on a free account. What differs by plan is what the nodes can then do, not whether you can connect.
Two operations need more than this token offers, and both say so on the node. Capsule: Update User Preferences requires the user_preference scope, which a personal access token does not document, so a 403 there is the expected answer. Capsule: Subscribe REST Hook and Capsule: Unsubscribe REST Hook require a registered application authenticated through the full OAuth flow plus an Administrator role, so a personal access token can list an account's hooks but cannot create one.

Connect Capsule CRM to Mattering

In Mattering Automations, open Connectors and click Add Connector.
Choose Capsule CRM from the list.
Click Create new credential and paste the personal access token into the bearer token field, then save it.
The token is stored encrypted in Google Secret Manager and never appears in workflow logs. Deleting it in Capsule ends the connection immediately, and the nodes then answer 401.
Click Test Connection. Mattering calls Capsule's current user endpoint, which takes no parameters, costs nothing on a large account and cannot be confused by an empty one, so a successful response proves the token and the header wiring together.
A 401 and a 403 are different findings here. 401 means the token is invalid, expired or mistyped. 403 means the token is valid and its scope is too narrow for that call, which is a scope problem rather than a credential problem. The connection test says which one it got.
All 131 Capsule nodes appear in your canvas palette.

Verify it works

Open any automation canvas and check the palette: a Capsule CRM section should list the read and action nodes. Drop Capsule: Get Current User onto the canvas and run a test step. It is the integration's natural probe, because it takes no parameters and answers the question that matters before any write, which is whose access this connection carries. Every node runs as that user, so if a later read returns fewer records than you expect, this is where the explanation usually starts.

Now run Capsule: List Parties without changing anything. Look at the type field on the rows that come back rather than only at the count: people and organisations are returned interleaved in one array, a person row carries firstName and lastName with no name, and an organisation row carries name with neither of the others. That is the single most important thing to see with your own eyes before you wire a template to a field, because a template pointed at the wrong one produces nothing for half the account and reports success while it does it.

Open a picker on a node such as the milestone field on Capsule: Create Opportunity and confirm it populates with milestone names grouped under their pipeline. Then look at Capsule: List Teams. If it comes back empty, check your plan before you check the connection: the free and Starter plans include zero teams, so an empty list there is a plan answer rather than a fault. Everything in this verify path is read-only, so it cannot reach any of the 16 delete nodes and is safe to re-run whenever you want to check the connection.

Node reference

131 nodes, generated from the connector itself so this reference always matches what ships.

Read and query nodes 67

Capsule: Get Account Site

capsule_get_site

Show which Capsule account the connected token belongs to: its subdomain, its display name and its URL. Takes no parameters. Useful for building links back into Capsule from a notification and for confirming an automation is pointed at the account somebody meant.

Capsule: Get Activity Type

capsule_get_activity_type

Fetch one activity type by id, with its icon and its updateLastContacted flag.

Capsule: Get Current User

capsule_get_current_user

Show the user the connected token belongs to. Two uses: it is the cheapest way to confirm a credential works, and it tells an automation whose identity its writes will be attributed to when no owner is set, which for a shared connector credential is worth knowing before creating records.

Capsule: Get Custom Field

capsule_get_field_definition

Fetch one custom field definition, including its options for a list type and its captureRule, which on party fields limits the field to people or to organisations only.

Capsule: Get Entry

capsule_get_entry

Fetch one history entry by id. Branch on data.entry.type before reading it: note, email and task entries carry different fields.

Capsule: Get Goal

capsule_get_goal

Fetch one goal by id, with its type, frequency, target value, date range and assignee.

Capsule: Get Lost Reason

capsule_get_lost_reason

Fetch one lost reason by id, with its includedForConversion flag.

Capsule: Get Milestone

capsule_get_milestone

Fetch one milestone by id, with its pipeline, probability, complete flag and stale threshold.

Capsule: Get Multiple Opportunities

capsule_get_opportunities_bulk

Fetch up to ten opportunities in one request by passing their ids comma separated. Capsule answers 404 for the whole request if any single id is unknown.

Capsule: Get Multiple Parties

capsule_get_parties_bulk

Fetch up to ten parties in one request by passing their ids comma separated. Worth using inside a loop that would otherwise spend ten of the account's 4,000 hourly requests. Capsule answers 404 for the WHOLE request if any one id is unknown, so it is a batch read for trusted ids rather than an existence check.

Capsule: Get Multiple Pipelines

capsule_get_pipelines_bulk

Fetch up to ten pipelines in one request by passing their ids comma separated. Capsule answers 404 for the whole request if any single id is unknown.

Capsule: Get Multiple Projects

capsule_get_projects_bulk

Fetch up to ten projects in one request by passing their ids comma separated. Capsule answers 404 for the whole request if any single id is unknown.

Capsule: Get Opportunity

capsule_get_opportunity

Fetch one opportunity by id. The read only lastStageChangedAt and lastOpenMilestone fields are the ones worth knowing about: they answer how long has this sat where it is without keeping a history table on our side.

Capsule: Get Party

capsule_get_party

Fetch one party by id. The record may be a person or an organisation and the payload differs accordingly, so branch on data.party.type before reading a name. Answers 404 when no party has that id.

Capsule: Get Pipeline

capsule_get_pipeline

Fetch one pipeline by id. The payload is just id, name and description; the milestones that make up the pipeline are a separate read, Capsule: List Milestones for Pipeline.

Capsule: Get Project

capsule_get_project

Fetch one project by id. status is OPEN or CLOSED and closedOn is read only, set by Capsule when the project closes rather than by an update.

Capsule: Get Project Board

capsule_get_board

Fetch one project board by id. The payload is id, name and description; its stages are a separate read, Capsule: List Stages for Board.

Capsule: Get Project Stage

capsule_get_stage

Fetch one project stage by id, with the board it belongs to and its position in that board's sequence.

Capsule: Get REST Hook

capsule_get_rest_hook

Fetch one webhook subscription by id. Capsule makes a hook visible only to the registered application that subscribed it, so a different connection answers as though it does not exist. This manages a webhook INSIDE Capsule. It is not a Mattering trigger and it does not make an automation start when something happens in Capsule; the target URL is a receiver you already run.

Capsule: Get Tag Definition

capsule_get_tag

Fetch one tag definition. dataTag tells you whether the tag carries custom fields of its own.

Capsule: Get Task

capsule_get_task

Fetch one task by id. hasTrack tells you whether the task was created by a track rather than by hand, which matters before deleting it: removing one task from a track leaves a gap in a sequence somebody designed.

Capsule: Get Task Category

capsule_get_category

Fetch one task category by id, with its colour.

Capsule: Get Team

capsule_get_team

Fetch one team by id. Ask for the memberships embed to list its members.

Capsule: Get Track

capsule_get_track

Fetch one applied track by id, with its generated tasks. direction is START_DATE, meaning trackDateOn is the anchor the task dates were counted forward from.

Capsule: Get Track Definition

capsule_get_track_definition

Fetch one track template with its full ordered list of task definitions, their day offsets and their assignee rules.

Capsule: Get User

capsule_get_user

Fetch one user by id. Ask for the party embed to reach their email address and phone number, which live on the linked party record rather than on the user.

Capsule: List Activity Type Icons

capsule_list_activity_type_icons

List the icons an activity type can use. A fixed set defined by Capsule, the same on every account, so it never changes per customer. Takes no parameters. This exists mainly so the icon field on the activity type nodes can offer real choices instead of asking for a magic string.

Capsule: List Activity Types

capsule_list_activity_types

List the activity types on the account. An activity type categorises a HISTORY ENTRY, which is what makes a note show as a Call or a Meeting on a record's timeline and what activity reporting counts. Capsule's built in types have NEGATIVE ids (Note is -1, Task completed is -4, Email sent is -5); custom types added on a paid plan have positive ones. This is the lister behind the activity type picker on Capsule: Create Note.

Capsule: List Additional Parties on Opportunity

capsule_list_opportunity_parties

List the extra contacts attached to an opportunity, which is how Capsule models a deal with several stakeholders. The MAIN party is not in this list; it is the party field on the opportunity itself. Rows are polymorphic parties like every other party read.

Capsule: List Additional Parties on Project

capsule_list_project_parties

List the extra contacts attached to a project. The MAIN party is not in this list; it is the party field on the project itself. Rows are polymorphic parties.

Capsule: List Countries

capsule_list_countries

List the country names Capsule accepts in a party's address, each with its ISO alpha-2, alpha-3 and numeric codes. The same list on every account. Worth calling before an import, because the address country field wants Capsule's NAME for the country rather than a code.

Capsule: List Currencies

capsule_list_currencies

List the ISO 4217 currency codes Capsule supports, with their symbols. The same list on every account. This is the lister behind every currency picker in the connector: opportunity value, goal target and user preference.

Capsule: List Custom Fields

capsule_list_field_definitions

List the custom field DEFINITIONS for one record type. This is the schema, not the values: a definition has a name, a type and for list types its options, and it may belong to a DataTag, in which case the fields only appear on records carrying that tag. Values live on the records and need the fields embed to come back at all. Definition ids are scoped per record type, so the same id can mean different fields under parties and under opportunities.

Capsule: List Custom Titles

capsule_list_custom_titles

List the CUSTOM person titles added to this account, such as Dr or Sir. Read the name carefully: this returns only what the account has added and NOT Capsule's built in titles, so a person whose title is Mr will not have that value appear here. That is why the title field on Capsule: Create Person is free text rather than a picker.

Capsule: List Deleted Opportunities

capsule_list_deleted_opportunities

List opportunities deleted since a date, as id and deletedAt pairs, plus a restrictedOpportunities array for records hidden from this token rather than removed. The companion to List Opportunities with since when keeping a local copy in step. Takes no page or perPage.

Capsule: List Deleted Parties

capsule_list_deleted_parties

List parties deleted since a date, as bare id and deletedAt pairs. This is the other half of a Capsule sync: List Parties with since tells you what changed, and this tells you what vanished, which no amount of re-reading the live collection will. The response also carries restrictedParties when records were made invisible to this token rather than deleted, and those look identical from the outside, so treat both as gone for this user. Note the endpoint takes no page or perPage.

Capsule: List Deleted Projects

capsule_list_deleted_projects

List projects deleted since a date, as id and deletedAt pairs. The companion to List Projects with since when keeping a local copy in step. Takes no page or perPage.

Capsule: List Entries

capsule_list_entries

Read the whole account's activity history, newest entry date first. Useful as an account-wide feed of what happened. Note what is NOT here: this endpoint takes no since parameter and no date range, unlike the party, opportunity and project listers, so an incremental sync of history means walking pages from the top and stopping at a known id.

Capsule: List Entries on Record

capsule_list_entity_entries

Read the history on one party, opportunity or project: notes, emails and the record of completed tasks, newest first. The type field on each row tells them apart and they are NOT the same shape. A note has content. An email also has subject and a participants array carrying FROM, TO, CC and BCC addresses, and it attaches to a parties array rather than a single party. A task row is the trace a completed task left behind. perPage is capped at 50 on this endpoint, not the usual 100.

Capsule: List Goal Periods

capsule_list_goal_periods

List the periods one goal is broken into, each with its target and its current value. This is where actual progress lives: the goal itself carries the target, and the periods carry what has been achieved. Note that a period's currentValue describes ONE period, so adding periods together across a goal is only meaningful for count and sum goal types and never for the targets.

Capsule: List Goals

capsule_list_goals

List the goals on the account. A goal is a target assigned to a user, measured against opportunities created, won or moved to a milestone, and broken into periods by its frequency. status is read only and is one of UPCOMING, ACTIVE, COMPLETED or MISSED. Goals are a paid-plan feature and the free plan has none.

Capsule: List Lost Reasons

capsule_list_lost_reasons

List the reasons an opportunity can be marked lost for. Capsule seeds every account with a set such as Price too high, Competition and Became Stale. The endpoint takes no parameters at all: no paging, no filter. This is the lister behind the lost reason picker on the opportunity nodes.

Capsule: List Milestones

capsule_list_milestones

List every milestone on the account across all pipelines. Each row carries the pipeline it belongs to, its win probability and its daysUntilStale threshold. This is the lister behind the milestone picker on the opportunity nodes, where the options are grouped by pipeline name.

Capsule: List Milestones for Pipeline

capsule_list_pipeline_milestones

List the milestones belonging to one pipeline, which is the ordered set of steps an opportunity in that pipeline moves through.

Capsule: List Opportunities

capsule_list_opportunities

List the opportunities on the account, which is Capsule's word for a potential sale: a bid, a deal or a proposal. Every opportunity has a main party, and that party may be a person or an organisation, so data.opportunities[].party.name is filled for company deals and empty for individuals. Pass since to fetch only what changed. Tags and custom field values need embed.

Capsule: List Opportunities for Party

capsule_list_party_opportunities

List every opportunity attached to one party. Covers deals where the party is the MAIN contact; a party added as an additional contact does not show up here, which is a real gap when a deal spans several stakeholders. There is no since parameter on this variant.

Capsule: List Parties

capsule_list_parties

List the people AND organisations on the Capsule account. Both kinds come back interleaved in one parties array, told apart by the type field: a person carries firstName, lastName, title and jobTitle, an organisation carries name, and neither carries the other. Wire a template to the wrong one and it silently produces nothing for half the records. Pass since to fetch only what changed, which is what Capsule recommends instead of re-reading the account. Tags and custom field values are NOT included unless embed asks for them. This is also the connector's auth probe: an invalid token answers 401 here.

Capsule: List People at Organisation

capsule_list_organisation_people

List the people linked to one organisation as employees. The id must be an ORGANISATION party; pointing it at a person returns nothing useful, since Capsule models the link in one direction only. Every row here is a person, which makes this the one party-returning read whose payload is not polymorphic.

Capsule: List Pipelines

capsule_list_pipelines

List the sales pipelines on the account. A pipeline is an ordered set of milestones an opportunity moves through. Every account has at least one and the free plan is capped at exactly one. Archived pipelines are hidden unless include_deleted asks for them.

Capsule: List Project Boards

capsule_list_boards

List the project boards on the account. A board is the project-side equivalent of a sales pipeline: a visual process made of stages that a project moves through from planning to completion. An account may hold at most 30 active boards, and the free plan has none at all.

Capsule: List Project Stages

capsule_list_stages

List every project stage on the account across all boards. Each row carries the board it belongs to and its displayOrder within that board. This is the lister behind the stage picker on the project nodes, where the options are grouped by board name.

Capsule: List Projects

capsule_list_projects

List projects, which is Capsule's word for work tracked after the sale: onboarding, delivery, a support case, an event. The response key is kases and the path is /kases because Capsule renamed Cases to Projects in the product and deliberately did not rename the API. Every project has a main party and an optional link back to the opportunity it came from. Pass since to fetch only what changed.

Capsule: List Projects for Opportunity

capsule_list_opportunity_projects

List the projects created to deliver an opportunity, which is the won-deal-to-delivery link. The path says kases because Capsule renamed Cases to Projects in the product and left the API alone, and the response key is kases for the same reason.

Capsule: List Projects for Party

capsule_list_party_projects

List every project whose MAIN party is this person or organisation. Projects where the party is only an additional contact are not included. No since parameter on this variant.

Capsule: List REST Hooks

capsule_list_rest_hooks

List the webhook subscriptions registered on the Capsule account, each with its event, its target URL and whether deliveries are batched. This manages a webhook INSIDE Capsule. It is not a Mattering trigger and it does not make an automation start when something happens in Capsule; the target URL is a receiver you already run. When the connection authenticates as a registered application, Capsule returns only the hooks THAT application subscribed, so an empty list can mean either no hooks or none of yours. Takes no parameters.

Capsule: List Stages for Board

capsule_list_board_stages

List the stages on one project board, in the order projects move through them. Choose the board first: this is the dependent half of the board and stage pair.

Capsule: List Tag Definitions

capsule_list_tags

List the tags DEFINED on the account for one record type. Tags are per record type in Capsule, so party tags, opportunity tags and project tags are three separate vocabularies and the entity field chooses which one. This lists the vocabulary, not which records carry which tag; to read a record's tags ask for the tags embed on the record itself, and to change them use the tags field on Update Party, Update Opportunity or Update Project.

Capsule: List Task Categories

capsule_list_categories

List the task categories on the account, the coloured labels used to group and report on tasks: Call, Email, Meeting and whatever the account has added. This is the lister behind the category picker on the task nodes. Do not confuse these with activity types, which categorise HISTORY ENTRIES rather than tasks.

Capsule: List Tasks

capsule_list_tasks

List tasks on the account. By default Capsule returns only OPEN tasks, which is the single most surprising thing about this endpoint: a workflow looking for completed work gets an empty list and no error. Pass status to widen it. There is no since parameter and no date filter here, so date narrowing means either paging and filtering downstream or using Capsule: Run Filter Query.

Capsule: List Teams

capsule_list_teams

List the teams on the account. Teams own records alongside or instead of individual users, which is how Capsule models shared ownership. Ask for the memberships embed to see who is in each one. Teams are a paid-plan feature starting on Growth, so this returns an empty list on a free or starter account rather than an error. Capsule offers no way to create or edit a team over the API; these two read nodes are the whole surface.

Capsule: List Track Definitions

capsule_list_track_definitions

List the track TEMPLATES on the account, each with the task definitions it will generate. A definition's captureRule says which record type it may be applied to, so filter by type when an automation is about to apply one. This is the lister behind the definition picker on Capsule: Apply Track.

Capsule: List Tracks on Record

capsule_list_entity_tracks

List the tracks running on one party, opportunity or project. A track is a track definition that has been APPLIED, so it holds the real dated tasks Capsule generated from the template. Ask for the task embed to see them.

Capsule: List Users

capsule_list_users

List the users on the Capsule account, with their status, timezone, locale and preferred currency. Each user also has a PARTY record holding their name, email addresses and phone numbers; ask for the party embed to get it. This is the lister behind every owner and assignee picker in the connector, and it takes no paging parameters.

Capsule: Run Filter Query

capsule_run_filter

Run a structured query over parties, opportunities or projects and get back matching records. This is the one node that answers questions the search and list nodes cannot: everyone tagged VIP, every deal added in the last seven days, every project whose custom field has a given value. It is a POST because the query travels in the body, but it READS ONLY and changes nothing. Results come back in exactly the shape the matching list node returns, so parties queries are polymorphic like every other party read.

Capsule: Search Opportunities

capsule_search_opportunities

Free text search over opportunities, matching what the global search box inside Capsule returns. For a structured query on a named field, a tag or a date range, use Capsule: Run Filter Query with entity set to opportunities.

Capsule: Search Parties

capsule_search_parties

Free text search across people and organisations, returning what the global search box inside Capsule would. Matches names, postcodes and phone numbers loosely, so it is the right tool for does this contact already exist and the wrong one for give me everyone tagged VIP. Results are polymorphic like every other party read.

Capsule: Search Projects

capsule_search_projects

Free text search over projects, returning what the global search box inside Capsule would. For a structured query on a named field, a tag or a date range, use Capsule: Run Filter Query with entity set to kases.

Action nodes 64

Capsule: Add Party to Opportunity

capsule_add_opportunity_party

Attach a person or organisation to an opportunity as an additional contact. Both ids go in the URL and there is no request body. Answers 204 with an empty body on success. Reversible with Capsule: Remove Party from Opportunity.

Capsule: Add Party to Project

capsule_add_project_party

Attach a person or organisation to a project as an additional contact. Both ids go in the URL and there is no request body. Answers 204 with an empty body.

Capsule: Apply Track

capsule_apply_track

Apply a track definition to a party, opportunity or project, which CREATES all of the definition's tasks at once with dates counted from the track date. This is the closest thing Capsule has to a workflow trigger and it is the highest-leverage write in the connector: one call can schedule an entire onboarding or chase sequence. Set exactly one of party, opportunity or kase, and note these three take a BARE numeric id here rather than an object. A track definition carries a capture rule saying which record types it may be applied to.

Capsule: Archive Pipeline

capsule_archive_pipeline

Archive a pipeline. Capsule spells this as a DELETE but calls it archiving, and it is REVERSIBLE with Capsule: Restore Pipeline, which is why it is not marked destructive. An archived pipeline stops being offered for new opportunities and is hidden from List Pipelines unless include_deleted asks for it.

Capsule: Archive Project Board

capsule_archive_board

Archive a project board. Capsule spells this as a DELETE but calls it archiving, and it is REVERSIBLE with Capsule: Restore Project Board, which is why it is not marked destructive. An archived board stops being offered and drops out of List Project Boards unless status asks for archived or all. It also frees a slot against the 30 active board cap.

Capsule: Create Activity Type

capsule_create_activity_type

Add a custom activity type. Capsule states this is only possible on a teams edition account, and its plan table shows custom activity types starting on the Growth plan, so a free or starter account gets a plan error rather than an auth error. The icon is chosen from Capsule's fixed set; Capsule: List Activity Type Icons returns it.

Capsule: Create Custom Field

capsule_create_field_definition

Define a new custom field on a record type. This is account configuration: it creates the field everywhere, it does not set a value anywhere. Setting values is the fields array on Update Party, Update Opportunity or Update Project. The free plan caps the account at 5 custom fields in total.

Capsule: Create Custom Title

capsule_create_custom_title

Add a custom person title to the account. An account may hold at most 50, and the operation requires the token's user to hold the Administrator role. Note the body root here is personTitle, not title, which the runner handles.

Capsule: Create Goal

capsule_create_goal

Create a goal for one user. Capsule requires a name, a goal type, a frequency, a target value and a date range, and the currency is required for the two SUM goal types and must be null for the COUNT ones. Scope it with at most one of pipeline or milestone, never both. Goals are a paid-plan feature.

Capsule: Create Lost Reason

capsule_create_lost_reason

Add a reason to the account's lost-reason list. Account configuration, not per-deal data: recording that a specific opportunity was lost is done by updating the opportunity. Requires the token's user to hold the Administrator role.

Capsule: Create Milestone

capsule_create_milestone

Add a milestone to a pipeline. probability is the default win chance Capsule prefills on opportunities that reach this step, and daysUntilStale is how long an opportunity may sit here before Capsule flags it as stale. Capsule's own create example omits the pipeline entirely, which puts the milestone on the account's default pipeline; name one explicitly whenever the account has more than one. Customizable sales milestones are a paid-plan feature, so this answers a plan error on the free plan.

Capsule: Create Note

capsule_create_note

Write a note onto a party, an opportunity or a project. The node is named for what the endpoint can actually do: Capsule states that type must be note and that note is currently the only accepted option, so this cannot post an email into the history. Set EXACTLY ONE of party, opportunity or project; Capsule requires exactly one and rejects zero or two. Choosing an activity type is what makes the note show as a Call or a Meeting rather than a plain Note, and some activity types also update the record's lastContactedAt.

Capsule: Create Opportunity

capsule_create_opportunity

Create an opportunity. Capsule requires three things: a name, a party to attach it to, and a milestone. The milestone is what decides which pipeline the deal lives in, so there is no separate pipeline field. An owner or a team is also required in practice; leaving both empty makes the API token's own user the owner, which for a connector is a shared service identity. Returns the created opportunity with its new id.

Capsule: Create Organisation

capsule_create_organisation

Create an ORGANISATION. This is POST /parties with type pinned to organisation, so name is the only identity field and none of the person fields appear. Use Capsule: Create Person for an individual. name is required by Capsule and a create without it answers 422. Returns the created organisation with its new id.

Capsule: Create Person

capsule_create_person

Create a PERSON. This is POST /parties with type pinned to person, so the fields on offer are the ones a person actually has. Use Capsule: Create Organisation for a company. Linking to an employer is done through the organisation field, and naming an organisation that does not exist yet CREATES it in the same call, which is convenient and is also how duplicate companies appear. Capsule runs no duplicate check here, so calling this twice with the same email makes two people; search first if the workflow can retry. Returns the created person with its new id.

Capsule: Create Pipeline

capsule_create_pipeline

Create a sales pipeline. A new pipeline needs milestones before opportunities can use it, so pair this with Capsule: Create Milestone, or use Capsule: Duplicate Pipeline to copy an existing one's milestones. Multiple pipelines are a paid-plan feature: the free and starter plans allow exactly one, so this node answers a plan error there rather than an auth error.

Capsule: Create Project

capsule_create_project

Create a project. Capsule requires a name and a main party. Setting a stage is what places the project on a board, so there is no separate board field; accounts on the free plan have no boards and therefore no stages to choose. Linking an opportunity records that this delivery came from that deal, which is the natural next step after a win. Returns the created project with its new id.

Capsule: Create Project Board

capsule_create_board

Create a project board. A new board needs stages before projects can move along it, so pair this with Capsule: Create Stage. Capsule caps an account at 30 active boards and the operation requires the token's user to hold the Administrator role, so a non-admin token answers 403 rather than failing validation.

Capsule: Create Project Stage

capsule_create_stage

Add a stage to a project board. Note the inconsistency worth knowing about: board is sent here as a BARE NUMERIC ID, not as an object, which is different from how owner, party and milestone are sent elsewhere in this API. Requires the Administrator role. The free plan has no boards, so there is nothing to add a stage to there.

Capsule: Create Tag Definition

capsule_create_tag

Define a new tag for one record type. This adds it to the account's vocabulary; it does not tag anything. Applying it to a record is done through the tags field on the update nodes, which accept a tag by id or by name.

Capsule: Create Task

capsule_create_task

Create a task, which in Capsule is a reminder of something to do. description is the one line title and is required; detail is the longer body. A task may hang off a party, an opportunity or a project. Set as many of the three as apply; leaving all three empty makes a standalone reminder. Returns the created task with its new id.

Capsule: Create Task Category

capsule_create_category

Add a task category to the account. Account configuration: it creates the label, it does not categorise any task. Assigning one is the category field on Capsule: Create Task or Capsule: Update Task.

Capsule: Create Track Definition

capsule_create_track_definition

Create a track template: a named sequence of tasks with day offsets, which can then be applied to records with Capsule: Apply Track. This is account configuration rather than day-to-day data, so it is the kind of node a setup automation runs once rather than one a workflow calls repeatedly.

Capsule: Delete Activity Type

capsule_delete_activity_type

PERMANENTLY remove an activity type from the account. Capsule documents no undelete and does not document what happens to existing history entries carrying the type, which is the reason to be careful: activity reporting is built on these. Answers 204 with an empty body.

Capsule: Delete Custom Field

capsule_delete_field_definition

PERMANENTLY delete a custom field definition AND every value stored in it, across every party, opportunity and project. Capsule documents no undelete. This is the most quietly expensive delete in the connector: the records survive, so nothing looks broken, and a column of data is simply gone. Answers 204, or 202 Accepted when Capsule queues the work, which is a SUCCESS and must not be retried.

Capsule: Delete Custom Title

capsule_delete_custom_title

Delete a custom person title. Capsule REFUSES while any party still uses the title, which makes this the safest delete in the connector: it cannot silently strip data, it just fails until the title is unused. Requires the Administrator role. Answers 204 with an empty body. Capsule documents no way to UPDATE a title, so renaming means delete then create.

Capsule: Delete Entry

capsule_delete_entry

PERMANENTLY delete a history entry AND its attachments. Capsule documents no undelete, and the attachments go with it, so this can destroy the only copy of a file. Answers 204 with an empty body, or 202 Accepted when Capsule queues the work, which is a SUCCESS and must not be retried.

Capsule: Delete Goal

capsule_delete_goal

PERMANENTLY delete a goal and its periods. Capsule documents no undelete, and the recorded progress against every past period goes with it. Answers 204 with an empty body.

Capsule: Delete Lost Reason

capsule_delete_lost_reason

Remove a lost reason from the account. Softer than most deletes here: Capsule states it MAY KEEP the reason for reporting on historical records while refusing to offer it on new or updated opportunities. It is still marked destructive because there is no documented undo and the account loses a reporting category. Answers 204 with an empty body. Requires the token's user to hold the Administrator role.

Capsule: Delete Milestone

capsule_delete_milestone

PERMANENTLY delete a milestone from its pipeline. Capsule documents no undelete and does not document what happens to opportunities currently sitting at that milestone, which is reason enough to move them first. Answers 204 with an empty body.

Capsule: Delete Opportunity

capsule_delete_opportunity

PERMANENTLY delete an opportunity. Capsule documents no undelete. If the intent is to record a deal that did not happen, update it with a closed date and a lost reason instead, which keeps it in reporting. Answers 204 with an empty body.

Capsule: Delete Party

capsule_delete_party

PERMANENTLY delete a party. Capsule documents no undelete and no trash for parties. Everything hanging off the record goes with it. Answers 204 with an empty body, or 202 Accepted when Capsule queues the deletion for later, which is a SUCCESS and must not be retried. Deleted ids show up afterwards in Capsule: List Deleted Parties.

Capsule: Delete Project

capsule_delete_project

PERMANENTLY delete a project. Capsule documents no undelete. To retire a project while keeping its history, update its status to CLOSED instead. Answers 204 with an empty body.

Capsule: Delete Project Stage

capsule_delete_stage

PERMANENTLY remove a stage from the account. Capsule documents no undelete and does not document what happens to projects currently sitting on that stage, which is reason enough to move them first. Answers 204 with an empty body. Requires the Administrator role.

Capsule: Delete Tag Definition

capsule_delete_tag

PERMANENTLY delete a tag definition AND remove it from every record carrying it. Capsule documents no undelete, so a tag applied to a thousand contacts takes that segmentation with it. Answers 204, or 202 Accepted when Capsule queues the work, which is a SUCCESS and must not be retried.

Capsule: Delete Task

capsule_delete_task

PERMANENTLY delete a task. Capsule documents no undelete. To record that the work was done, update the task to COMPLETED instead: that keeps it in activity reporting and writes a completed-task entry onto the record's history, which deleting does not. Check hasTrack first, because deleting a track task breaks a designed sequence. Answers 204 with an empty body.

Capsule: Delete Task Category

capsule_delete_category

Remove a task category from the account. The gentlest delete here: Capsule keeps every task and simply leaves them uncategorised. Still marked destructive because the categorisation itself is gone from those tasks with no documented undo. Answers 204 with an empty body.

Capsule: Delete Track

capsule_delete_track

Remove an applied track from its record AND DELETE EVERY TASK IT CREATED. That second half is the reason this is destructive rather than a tidy-up: a track that generated ten tasks takes all ten with it, including any a person has already been working on. Capsule documents no undelete. Answers 204 with an empty body.

Capsule: Delete Track Definition

capsule_delete_track_definition

PERMANENTLY delete a track template. Capsule documents no undelete. Deleting the TEMPLATE is not the same as deleting the tracks applied from it, which are separate records with their own tasks. Answers 204 with an empty body.

Capsule: Duplicate Pipeline

capsule_duplicate_pipeline

Copy an existing pipeline, milestones included, which is the quick way to stand up a second process that mostly matches the first. No request body. The copy is a new pipeline with new ids; existing opportunities are not moved into it. Subject to the same paid-plan pipeline limit as create.

Capsule: Remove Party from Opportunity

capsule_remove_opportunity_party

Detach an additional contact from an opportunity. This removes the LINK only; neither the party nor the opportunity is deleted, which is why it counts as a reversible write rather than a destructive one. Answers 204 with an empty body.

Capsule: Remove Party from Project

capsule_remove_project_party

Detach an additional contact from a project. Removes the LINK only; neither record is deleted. Answers 204 with an empty body.

Capsule: Restore Pipeline

capsule_restore_pipeline

Bring an archived pipeline back into use. No request body. Find the id with Capsule: List Pipelines and include_deleted set true, since an archived pipeline is otherwise not listed.

Capsule: Restore Project Board

capsule_restore_board

Bring an archived project board back into use. No request body. Find the id with Capsule: List Project Boards and status set to archived, since the default picker only shows active boards.

Capsule: Subscribe REST Hook

capsule_subscribe_rest_hook

Subscribe a webhook so Capsule POSTs to a URL when an event happens. This manages a webhook INSIDE Capsule. It is not a Mattering trigger and it does not make an automation start when something happens in Capsule; the target URL is a receiver you already run. REQUIRES OAUTH, NOT A PERSONAL ACCESS TOKEN. Capsule states that only registered applications can manage subscriptions and that the caller must be authenticated through the full OAuth flow, and separately that the user must hold the Administrator role. A connection holding a personal access token will not be able to run this node. An account may hold 20 subscriptions at once. Delivery is retried five times over 36 minutes and 5 seconds, after which Capsule marks the hook dead and unsubscribes it, and a target answering 410 Gone is unsubscribed immediately, so an endpoint that goes down long enough disappears without further warning. Reversible with Capsule: Unsubscribe REST Hook.

Capsule: Unsubscribe REST Hook

capsule_unsubscribe_rest_hook

Remove a webhook subscription, stopping deliveries immediately. This manages a webhook INSIDE Capsule. It is not a Mattering trigger and it does not make an automation start when something happens in Capsule; the target URL is a receiver you already run. Reversible in the sense that Capsule: Subscribe REST Hook can recreate it, which is why it is not marked destructive; no Capsule data is deleted, only the subscription. REQUIRES OAUTH, NOT A PERSONAL ACCESS TOKEN. Capsule states that only registered applications can manage subscriptions and that the caller must be authenticated through the full OAuth flow, and separately that the user must hold the Administrator role. A connection holding a personal access token will not be able to run this node. Capsule further requires the SAME registered application that subscribed the hook. Answers 204 with an empty body.

Capsule: Update Activity Type

capsule_update_activity_type

Rename an activity type, change its icon, or change whether it moves lastContactedAt. Anything not sent stays as it is.

Capsule: Update Custom Field

capsule_update_field_definition

Partially update a custom field definition. Worth care with options on a list field: dropping a value that records already hold changes what those records mean, and Capsule does not warn about it here.

Capsule: Update Entry

capsule_update_entry

Edit the content of a note or an email in the history. Capsule restricts this endpoint to entries of type email or note, so it cannot rewrite the task rows. Only content and attachments are editable; the record the entry sits on, its activity type and its entry date are not. Removing the last attachment from an entry with no content is a validation error, so delete the entry instead.

Capsule: Update Goal

capsule_update_goal

Partially update a goal. Four things CANNOT be changed and are therefore not offered here: the assignee, the goal type, the pipeline and the milestone. Capsule states this outright, so a workflow that needs to move a goal to another user has to delete it and create a new one.

Capsule: Update Lost Reason

capsule_update_lost_reason

Rename a lost reason or change whether it counts toward the conversion rate. Anything not sent stays as it is. Requires the token's user to hold the Administrator role.

Capsule: Update Milestone

capsule_update_milestone

Partially update a milestone. Changing probability affects what Capsule prefills on opportunities reaching this step from now on, not the probability already stored on existing deals.

Capsule: Update Opportunity

capsule_update_opportunity

Partially update an opportunity: anything not sent stays as it is. Moving the deal along the pipeline means setting a new milestone, which is also what fires Capsule's own opportunity/moved webhook. Marking it lost means setting closedOn and a lost reason; the lost reason on its own is only a reporting label and closes nothing. An owner or a team is required on an opportunity, so do not clear both.

Capsule: Update Party

capsule_update_party

Partially update a party. One node covers both kinds because Capsule has one endpoint and no type check: fields you do not send stay unchanged. Send only the ones that belong to the kind the id already is. Person fields: first name, last name, title, job title, organisation. Organisation field: name. Everything else applies to both. The collections follow Capsule's three-way convention, so an entry with no id adds, an entry with an id edits, and an entry carrying _delete true removes. This node cannot change a person into an organisation.

Capsule: Update Pipeline

capsule_update_pipeline

Rename a pipeline or change its description. Its milestones are edited separately with the milestone nodes.

Capsule: Update Project

capsule_update_project

Partially update a project: anything not sent stays as it is. Moving it to a new stage is what fires Capsule's own kase/moved webhook, and setting status to CLOSED fires kase/closed and makes Capsule stamp the read only closedOn date. An owner or a team is required, so do not clear both.

Capsule: Update Project Board

capsule_update_board

Rename a project board or change its description. Anything not sent stays as it is. Requires the Administrator role.

Capsule: Update Project Stage

capsule_update_stage

Rename a project stage, change its description, or move it in the board's sequence with displayOrder. Capsule documents no way to move a stage to a DIFFERENT board here. Requires the Administrator role.

Capsule: Update Tag Definition

capsule_update_tag

Rename a tag or change its description, account wide. Turning a DataTag back into a plain tag has an order of operations: Capsule requires every custom field definition linked to the tag to be deleted FIRST, and only then accepts dataTag false.

Capsule: Update Task

capsule_update_task

Partially update a task: anything not sent stays as it is. Setting status to COMPLETED is how an automation ticks a task off, and it is what fires Capsule's own task/completed webhook. Completing a repeating task also makes Capsule create the next occurrence, readable through the nextTask embed.

Capsule: Update Task Category

capsule_update_category

Rename a task category or change its colour. Anything not sent stays as it is, and every task already carrying the category picks up the change.

Capsule: Update Track

capsule_update_track

Rename an applied track or change its anchor date. Capsule allows only these two attributes here; the track's tasks, its definition and the record it runs on cannot be changed. Whether moving the anchor date reschedules the already-generated tasks is not documented, so verify against a real account before relying on it.

Capsule: Update Track Definition

capsule_update_track_definition

Partially update a track template. Editing task_definitions uses Capsule's three-way convention, so a step with no id is added, a step with an id is changed, and a step carrying _delete true is removed from the template. Changing a template affects tracks applied AFTERWARDS; tracks already running were expanded into real tasks at apply time and do not change.

Capsule: Update User Preferences

capsule_update_user

Update a user's preferences: locale, currency, timezone, username, task reminder emails and the click behaviour for emails and phone numbers. THIS NODE NEEDS THE user_preference SCOPE. Capsule states it plainly, and the connector's ordinary read write token does not carry it, so a 403 here is the expected outcome rather than a broken connection. The user's name and contact details are not editable here; they live on the linked party record and are changed with Capsule: Update Party.

Using the integration

A party is not a contact, and the connector says so everywhere

Capsule has no contact type. It has a party, and a party is either a person or an organisation, told apart by a required type field. The two kinds do not share a name. Capsule's own party model states that firstName, lastName, title, jobTitle and organisation are present only when the type is person, and that name is present only when the type is organisation, so there is no single field that names both kinds. Three things follow, all deliberate. Creating is split into Capsule: Create Person and Capsule: Create Organisation, which are the same underlying endpoint with the type pinned, so each config panel offers the fields that kind actually has instead of ten fields half of which are always wrong. Reading and updating stay single nodes, because Capsule's update is a partial merge over whichever kind the id already is and performs no type check, so Capsule: Update Party labels its fields PERSON ONLY and ORGANISATION ONLY instead. And every party-returning list sample on the canvas carries one person row and one organisation row, so the draggable output paths in the Input column show both shapes rather than hiding half of your account behind a field that happens to be null in the sample.

Milestones belong to pipelines, stages belong to boards

These are two separate hierarchies and conflating them is the most common modelling mistake with Capsule. On the sales side, a pipeline is an ordered set of milestones and an opportunity sits on a milestone. The milestone is also what decides which pipeline a deal is in, which is why Capsule: Create Opportunity asks for a milestone and has no pipeline field: setting the milestone is the assignment. Capsule: List Milestones for Pipeline returns the ordered steps for one pipeline, each with the win probability Capsule prefills on deals that reach it and a stale threshold in days. On the project side, a board is an ordered set of stages and a project sits on a stage, so Capsule: Create Project takes a stage and has no board field for the same reason. Capsule: List Stages for Board is the dependent read there. The pickers in this integration group each one under its own parent, milestones by pipeline and stages by board, so the panel makes the hierarchy visible. If you find yourself looking for the stages on a sales pipeline, what you want is the milestone nodes.

Tags and custom field values are not in the response unless you ask

Capsule omits tags and custom field values from every party, opportunity and project by default and returns them only when the request carries an embed asking for them. This is the second most common cause of a workflow that runs green and produces nothing: the record came back, the record looked complete, and the custom field the next step wanted was simply not in the payload. Every node that supports it carries an embed field listing only the values Capsule documents for that endpoint, because an unsupported value is silently ignored rather than rejected. Those fields are deliberately optional and have no default, which is the opposite of how some connectors handle this: Capsule's default payload is already a usable record, so defaulting the embed would spend the account's hourly request budget widening every read whether the caller wanted it or not. Set it on the nodes that need it. Note also that a custom field definition and a stored custom field value are different things: Capsule: List Custom Fields returns the schema for a record type, not anybody's values, and setting a value is done through the record's own update node.

Incremental sync, and the caveat that comes with it

Capsule allows each user 4,000 requests per hour, and its own advice for keeping a local copy in step is the since parameter plus the deleted endpoints rather than repeated full reads. This integration encodes exactly that. Capsule: List Parties, Capsule: List Opportunities and Capsule: List Projects take a since date and return only what was added or updated after it. Capsule: List Deleted Parties, Capsule: List Deleted Opportunities and Capsule: List Deleted Projects return the ids that went away, which is the half most syncs skip and then keep stale rows forever. Two honest caveats. Capsule offers since and no until anywhere, so the window is one-sided: you can advance a scheduled refresh but you cannot split a backfill into partitions. And Capsule documents no sort order on those listers and pages by page number rather than by cursor, so a record updated while a workflow walks pages can move position and be missed. The practical mitigation is to overlap windows, re-requesting from slightly before the last high-water mark, and to treat the results as at-least-once rather than exactly-once.

REST Hooks manage Capsule's webhooks, they are not triggers here

The four REST Hook nodes are the part of this integration most likely to be misread, so it is worth being blunt. They manage subscriptions inside Capsule. A subscription registers a target URL that Capsule will POST to when something happens, and the events available are party created, updated and deleted, project created, updated, deleted, closed and moved, opportunity created, updated, deleted, closed and moved, task created, updated and completed, and user created, updated and deleted. None of these nodes makes a Mattering automation start when something changes in Capsule. They are account configuration in the same way a pipeline or a tag definition is, which is why they are read and action nodes rather than triggers. Capsule's own limits apply: an account may hold 20 subscriptions at once, hooks must be managed by account administrators, and only a registered application authenticated through the full OAuth flow can subscribe or unsubscribe one, so a personal access token can list hooks through Capsule: List REST Hooks but cannot create one. Capsule retries a failed delivery five times over roughly 36 minutes before marking the hook dead and unsubscribing it, and a target answering 410 Gone is unsubscribed immediately, which makes a scheduled read of the hook list a genuinely useful piece of hygiene. Until a Mattering trigger for Capsule exists, run Capsule workflows on a schedule, which the since parameter makes cheap, or start them from a trigger elsewhere in the platform.

Projects live at kases, and other names Capsule kept

Capsule renamed Cases to Projects in the product and deliberately did not rename the API, in its own words because that would be a breaking change for existing applications. So the underlying path is kases, the response key is kases, the write body root is kase, and the entity keyword you pass to the nodes that work across record types is kases. This integration says PROJECT in every label and description a human reads and uses the wire spelling only where a wire format demands it, so nothing in the palette says kase. The one place it surfaces is the entity field on the cross-type nodes, such as the tag, custom field, track and history-entry nodes, which accept parties, opportunities or kases. That field is validated before the URL is built, because a plausible typo like projects would otherwise produce a 404 from a path that looks entirely correct.

Writes, deletes, and the 202 that is not a failure

Capsule's updates are partial merges everywhere: fields you do not send stay unchanged, which makes Capsule: Update Opportunity and Capsule: Update Party safe to use for a single field. Inside collection fields such as addresses, phone numbers, tags and custom field values, Capsule uses a three-way convention that this integration passes through verbatim rather than hiding: an entry with no id is added, an entry with its id is updated, and an entry with its id plus a delete flag is removed. That is why those fields take JSON rather than being flattened, because flattening them would make the removal case impossible to express. On deletes, Capsule documents no undelete and no trash for any resource, so all 16 delete nodes are flagged destructive and several name a reversible alternative in their own description: complete a task rather than deleting it, close a project rather than deleting it, record a lost reason on an opportunity that did not happen. Two resources have a real softer path and get their own nodes: pipelines and boards archive and restore, through Capsule: Archive Pipeline with Capsule: Restore Pipeline and Capsule: Archive Project Board with Capsule: Restore Project Board. One response detail matters operationally: a delete normally answers 204 with an empty body, but Capsule may answer 202 Accepted when it queues the work as a long running operation, and party, history entry, tag and custom field deletes all document this. A 202 is a success in this integration's envelope, so a workflow that treats it as a failure will retry a delete that has already happened.

Pickers, and the one field that is deliberately free text

88 fields across 52 of the 131 nodes carry a picker rather than asking you to paste an id, sourced from the integration's own listers: users, teams, pipelines, milestones grouped by pipeline, boards, stages grouped by board, opportunities, projects, tasks, goals, categories, lost reasons, activity types and their icons, track definitions grouped by capture rule, currencies, custom titles, REST hooks, and both tag definitions and custom field definitions narrowed to whichever record type the node is pointed at. The deliberate exception is the party id field, which is the most used id in the integration and is free text on purpose. A picker reads one label path, and Capsule's party has no field that names both kinds, so labelling from name would render every person as a bare numeric id and labelling from lastName would do the same to every organisation. Roughly half of any account would show as numbers, which is the exact failure a picker exists to prevent, so the field takes an id and its description points at Capsule: Search Parties and Capsule: List Parties to find one. Three smaller fields are free text for their own stated reasons: a person's title, because the custom titles lister returns only what the account has added and never Capsule's built-in ones, a history entry id, because entry rows have no short human label, and a track id, because the lister needs an entity and an entity id that the single-track nodes do not carry.

On the automation canvas

Capsule nodes chain with every other node in Mattering. A form submission can run Capsule: Search Parties to avoid a duplicate, then Capsule: Create Person or Capsule: Create Organisation, then Capsule: Create Note so the history records where the contact came from. A scheduled workflow can read Capsule: List Opportunities with a since date, spot deals that closed as won, open the delivery record with Capsule: Create Project and schedule the whole onboarding in one call with Capsule: Apply Track, which creates every task in a track definition at once with dates counted from an anchor you pass. Capsule: Run Filter Query answers the structured questions the search endpoints cannot, such as everyone tagged VIP or every deal added in the last seven days, and it is a read, so it is safe to put in front of anything. Where a workflow would otherwise loop, prefer the batch reads: Capsule: Get Multiple Parties, Capsule: Get Multiple Opportunities, Capsule: Get Multiple Projects and Capsule: Get Multiple Pipelines each fetch up to ten records in one request.

From AI assistants over MCP

Every node speaks MCP, the open protocol for connecting AI assistants to tools, so Claude, ChatGPT, and Matter, the assistant built into the platform, can drive the same surface. Capsule does publish MCP servers of its own and they are worth knowing about: a local Docker server announced in Capsule's developer changelog on 15 April 2026 as a private beta with read-only access, its source public at github.com/zestia/capsule-mcp-scala, and a hosted connector published in Claude's connector directory that authenticates with your ordinary Capsule login. Two documented differences decide between them. Capsule states that read access lets the client view data in the CRM while write access is currently limited to updating Custom Fields, and its support page adds that Claude cannot rename a record, reassign its owner or touch any other core field, whereas this integration ships 64 action nodes that create and update parties, opportunities, projects, tasks, notes and account configuration. And Capsule lists its hosted connector as available on Growth, Advanced and Ultimate, while Capsule's API is on every plan including Free. Given that this is a CRM, the sensible split is to hand an assistant the 67 read nodes freely and keep the writes, particularly the 16 deletes, behind a condition or a human approval step, because nothing in Capsule has an undelete.

Troubleshooting

Test Connection fails with a 401
Capsule answers 401 when the bearer token is invalid or expired, and includes detail in the body and in a WWW-Authenticate header. In practice the causes are a mistyped or truncated paste, a token that was deleted from inside Capsule, or an OAuth access token that has aged out, since Capsule issues those for seven days and this integration does not refresh them for you. Generate a fresh personal access token from My Preferences then API Authentication Tokens and update the credential. A personal access token carries no documented expiry, so once it works it keeps working until somebody removes it.
A node fails with a 403 while other nodes work
A 403 is a different finding from a 401 rather than a worse one: Capsule uses it when the token is valid but does not carry enough scope for that request. Two cases account for nearly all of them here, and both are documented on the node. Capsule: Update User Preferences requires the user_preference scope, which a personal access token does not document, so a 403 there is the expected result rather than a defect. Capsule: Subscribe REST Hook and Capsule: Unsubscribe REST Hook require a registered application authenticated through the full OAuth flow, so a personal access token cannot create a subscription no matter whose it is. Beyond scope, several configuration writes require the token's user to hold the Administrator role, including boards, stages, lost reasons, custom titles and REST hooks, so check the acting user's role with Capsule: Get Current User before assuming the token is wrong.
A create fails with a message about a plan limit rather than a permission
This is Capsule enforcing your subscription and it is not a connector fault, which is why the message comes back with Capsule's own wording rather than as a 401 or a 403. The free plan caps contacts at 250 and custom fields at 5, and includes 1 sales pipeline, 0 project boards and 0 teams, with no goals, no customizable sales milestones and no custom activity types. So on a free account Capsule: Create Project Board, Capsule: Create Project Stage, Capsule: Create Goal, Capsule: Create Activity Type, a second Capsule: Create Pipeline and a sixth Capsule: Create Custom Field will all be refused while the party, opportunity, project, task, history, track and tag nodes work normally. Project boards start on Starter and teams start on Growth. Check the plan before you check the credential.
A picker is empty
Work out first whether the list is empty or the picker is unwired. Every picker is fed by a lister in this integration, so run that lister on its own: if Capsule: List Project Boards returns nothing, no stage picker will populate either, and on a free account it will return nothing because the plan includes zero boards. The same goes for Capsule: List Teams on Free and Starter. The second cause is a parent field that is not set yet, because several pickers are dependent: stages narrow to a chosen board, milestones to a chosen pipeline, and the tag and custom field pickers narrow to whichever record type the node's entity field names, so they stay empty until you pick parties, opportunities or kases. The party id field is free text by design rather than a broken picker, because Capsule's people and organisations have no shared name field and any single label would render half the account as bare ids.
A record came back but its custom fields or tags are missing
Capsule omits tags and custom field values from every party, opportunity and project unless the request asks for them, so the node worked and the payload genuinely did not contain them. Open the node's embed field and add the values you need, for example fields for custom field values and tags for tags. Each node lists only the embed values Capsule documents for that endpoint, which matters because Capsule silently ignores an unsupported value rather than rejecting it, so a typo there produces the same missing data with no error. One related distinction: Capsule: List Custom Fields returns field definitions for a record type, which is the schema, not anybody's stored values.
A task list looks empty or is missing tasks that exist
Capsule's task endpoint returns only OPEN tasks by default, which is the single most surprising thing about it. A workflow looking for completed work finds nothing and reports success. Set the status filter on Capsule: List Tasks to the state you actually want. The endpoint carries no date filter at all, so a date-bounded task report has to filter after reading rather than in the request. Completed tasks also appear in a record's history, so Capsule: List Entries on Record is the other place to look for work that was done.
I cannot find the stages on my sales pipeline
They do not exist, and this is a naming collision rather than a missing feature. Capsule has two separate hierarchies: milestones belong to pipelines on the sales side, and stages belong to boards on the project side. An opportunity sits on a milestone, which is also what decides its pipeline, so use Capsule: List Milestones or Capsule: List Milestones for Pipeline. A project sits on a stage, which is what places it on a board, so use Capsule: List Project Stages or Capsule: List Stages for Board. The two never mix, and no node takes a stage where a milestone belongs.
A list node returns fewer records than the account holds
Paging, almost always. Capsule numbers pages from 1, defaults to 50 records per page and caps a page at 100, except on the entries-for-a-record endpoint where the documented cap is 50. There is no cursor: to walk a collection you increment the page number. These nodes do not auto-paginate, deliberately, so that one canvas step is one request and an unbounded walk cannot hide inside it. If you are using a since date, note also that Capsule documents no sort order on those listers, so a record updated while you are walking pages can move position and be skipped. Overlap your windows and treat the results as at-least-once. The other possibility is access: every node runs as the token's user, so a record that user cannot see is not returned. Capsule: Get Current User confirms who that is.
A delete answered 202 and the workflow treated it as a failure
202 Accepted is a success here. Capsule normally answers a delete with 204 and an empty body, but for what it calls long running operations it may answer 202 with a Location header pointing at a job instead, meaning the request was accepted and is not being processed at this exact moment. Party, history entry, tag definition and custom field deletes all document this behaviour. This integration surfaces a 202 as a success, so no change is needed on the node, but a downstream condition written to expect exactly 204 will retry a delete that has already happened. Since Capsule documents no undelete for any resource, that retry is worth avoiding.
Requests start failing partway through a run with a 429
Capsule allows each user 4,000 requests per hour on bearer token authentication and returns a rate limit reached message at 429 once you pass it. Every successful response carries the limit, the number remaining and the epoch time the window resets, so wait for the reset rather than retrying immediately. The durable fix is shape rather than pacing, and it is Capsule's own advice: put a short delay between requests, replace repeated full reads with the since parameter on Capsule: List Parties, Capsule: List Opportunities and Capsule: List Projects plus the three deleted companions, and replace a loop of single reads with Capsule: Get Multiple Parties, Capsule: Get Multiple Opportunities, Capsule: Get Multiple Projects or Capsule: Get Multiple Pipelines, each of which fetches up to ten records in one request. Note that those batch nodes answer 404 for the whole request if any single id is unknown.
A template wired to a party field produces nothing for half the records
This is the polymorphism, and it is the most expensive mistake to make with Capsule because nothing fails. A party is either a person or an organisation and the two kinds have no shared name field: a person carries firstName and lastName and no name, an organisation carries name and neither of the others. A template pointed at one of them silently renders empty for every record of the other kind, at HTTP 200, with the workflow reporting success. Branch on the type field before reading a name, or build a fallback that takes name when it is present and the first and last name when it is not. The list samples on every party-returning node deliberately include one person row and one organisation row so both shapes appear in the canvas Input column.
A REST Hook node did not start my automation
It was never going to, and this is worth reading rather than debugging. The four REST Hook nodes manage webhook subscriptions inside Capsule: they register a target URL that Capsule posts to when a party, project, opportunity, task or user changes. They are Capsule account configuration, not Mattering triggers, and none of them makes an automation run when something happens in Capsule. Run Capsule workflows on a schedule instead, which the since parameter makes cheap, or start them from a trigger elsewhere in the platform such as a form submission or an inbound email. If you do want Capsule to post somewhere, remember the vendor limits: 20 subscriptions per account, administrators only, and only a registered application authenticated through the full OAuth flow can subscribe or unsubscribe, so a personal access token can list hooks but not create them.

Capsule CRM is a trademark of its respective owner. Mattering uses platform names to identify the services these integrations connect to; no endorsement or affiliation is implied.

Start Free Explore our Products