A Shortcut MCP that keeps working when nobody is watching
126 nodes on Shortcut REST API v3: stories, epics, iterations, objectives and key results, plus docs, teams, workflows, labels, categories, custom fields, story links, files, members and outgoing webhooks. 66 nodes read and 60 act. Use them from Claude, ChatGPT or any MCP-capable assistant, and use the same nodes as steps on a workflow that runs on a schedule with one credential shared across the workspace.
What the Shortcut MCP connector does
The Shortcut connector for Mattering Automations is 126 nodes built from Shortcut's own machine-readable OpenAPI document for REST API v3. Stories, sub-tasks, tasks, comments, reactions and story links cover the work item itself. Epics, epic comments, epic health and the epic workflow cover the layer above it. Iterations cover the sprint, and objectives, key results and objective health cover the strategic layer above that. Docs, files, linked files and labels cover the material hanging off the work, and teams, workflows, projects, categories, custom fields, members and repositories cover the reference data everything else points at. 66 nodes read and 60 act.
Every capability is a node on the Automations canvas, and every node speaks MCP, the open protocol for connecting AI assistants to tools. Claude, ChatGPT and any other MCP-capable assistant can drive them through your workspace, alongside Matter, the AI assistant built into the platform. The same nodes are canvas steps, so a question an assistant helps you answer once can become a workflow that answers it every Monday morning with nobody present.
Shortcut publishes its own MCP server, and it should be the first thing you weigh. It is remote, hosted at mcp.shortcut.com, and it authenticates with OAuth rather than an API token, so your client prompts you to sign in to Shortcut the first time you use it. Both of Shortcut's MCP repositories on GitHub are archived, with the README stating that all future development moved to the remote server, and Shortcut's pricing table lists MCP Server as included on every plan including Free. If what you want is a Shortcut-aware assistant in your editor, install theirs. This connector is a different shape of answer: it is the same Shortcut surface as workflow steps, so it runs on a schedule or on an inbound event without a person in the loop, chains with every other connector in one workflow rather than being the only tool in the conversation, holds one credential in a shared workspace store instead of in each person's local config, and carries a destructive flag on the nodes that delete.
Pagination on this API is genuinely heterogeneous, and it is worth knowing before you design anything, because a connector that pretended otherwise would be lying to you. Four different shapes exist on one API and every node declares which one it uses. 35 of the 70 read endpoints return a bare JSON array with no envelope, no page parameter, no cursor and no total, which means Shortcut: List Epics returns every epic the token can see in a single response. That is the default rather than the exception here. Exactly one path is page numbered, which is why Shortcut: List Epics (Paginated) exists for epics alone. The six search endpoints return a cursor that is not a token but the path and query string of the next page, and they cap out at 1000 total matches with a dedicated error rather than an empty page, so a loop that walks every page of a broad query ends in a failure rather than in exhaustion. And exactly one path, Shortcut: List Team Stories, takes limit and offset over a bare array with no total, so the only way to know you have reached the end is to receive fewer rows than you asked for.
Shortcut's documented rate limit is 200 requests per minute, answered with HTTP 429 past that, and its documentation specifies no Retry-After header on the response. That second half matters more than the number. A retry loop written against the usual header contract reads nothing, treats it as zero and hammers the limit it just hit, so back-off here has to be fixed or exponential rather than header driven. The connector surfaces the 429 as an ordinary unsuccessful result with that fact attached rather than retrying silently, so a workflow can branch on it. The durable fix is shape: on an API where most listers return the whole collection in one call, one wide read beats a loop of narrow ones.
Two facts about identity will save you a debugging session. A Shortcut API token acts as the member who minted it, which means anything that member cannot see in the web app is invisible to the connector, a 403 is usually that person's own role rather than a permission we can widen, and every comment written through Shortcut: Create Story Comment is attributed to them unless the node's author id names somebody else. And ids come in two kinds: stories, epics, iterations, objectives, labels, categories, projects, files, linked files and workflows use integer public ids, while members, teams, docs, custom fields, story templates and health records use UUID strings. Passing the wrong kind is the commonest 404 on this API, so every field on every node declares its type.
Deletes here are real deletes. 18 nodes delete, 16 of them are flagged destructive, every one answers 204 No Content, and Shortcut documents no undelete for any resource. The nodes say so individually rather than behind one blanket warning, because the blast radius differs: deleting an epic leaves its stories alive with their epic id cleared and nothing afterwards records which ones they were, deleting a label strips it from every story and epic carrying it so saved views filtered on it silently return nothing, and deleting an uploaded file takes the bytes with it and this connector cannot re-upload them. Where Shortcut supports archiving, the delete node points at it, because setting archived to true through the matching update node is the reversible operation people usually wanted. One delete has a built-in guard: Shortcut: Delete Stories in Bulk only removes stories that are already archived.
18 of Shortcut's 143 operations are deliberately not built, and the reasons are worth stating rather than hiding. Eight are milestone operations, six of which Shortcut's own document marks deprecated with the objective equivalent named in the description, because Shortcut renamed Milestones to Objectives and kept the old paths as aliases. Building both would ship two nodes writing to one place while disagreeing about what it is called. File upload is not built because it is a multipart form and a canvas has no local file to attach, so URL-shaped attachments go through the linked file nodes instead. The four workspace feature toggles that switch Iterations and Story Templates on or off for every member are not built, because that is an administrator's deliberate act in the web app rather than something an automation should reach by accident. Deleting a custom field definition is not built because it removes that field's values from every story in the workspace with no undelete, so the list, get and update nodes are there and the delete is not. Story template writes are not built because a template body is a whole nested document authored in Shortcut's own editor. Nothing is unreachable, though: Shortcut: API Request is an advanced node that calls any Shortcut v3 endpoint directly with your credential applied, including anything Shortcut ships after this connector was built.
One connection, three surfaces
- Visual automations. Drag Shortcut nodes onto the canvas and wire them to triggers, conditions, and other apps.
- AI agents, yours or ours. These nodes speak MCP, the open protocol, so Claude, ChatGPT, and any MCP-capable assistant can use them, and Matter, the agent built into the platform, uses them natively.
- The whole platform. Combine with native tasks, projects, and other connectors, no middleware in between.
Every Shortcut node, ready to drop on the canvas
126 nodes, generated from the connector itself so this list always matches what ships. Expand a category to browse.
Read and query nodes 66
Shortcut: Find Stories by External Link
Find every story carrying a given external link. This is the reverse lookup that makes a two-way sync possible: given a URL in another system, it answers which Shortcut stories point at it, without searching. The match is on the whole URL, so a trailing slash or a query string that differs will not match.
Shortcut: Get Category
Fetch one category with its name, colour and archived flag.
Shortcut: Get Current Member
Who this API token acts as: the member's id, name, mention name, role and owner flag, plus the workspace slug and its ESTIMATE SCALE. Two uses. It is the cheapest call on the API, so it is the natural connection test, and it is the only place the legal estimate values are published, which a story-creating automation needs before it sets an estimate.
Shortcut: Get Custom Field
Fetch one custom field definition with its type, enabled flag, the story types it applies to, and every value it allows.
Shortcut: Get Doc
Fetch one doc with its body. The response always carries content_markdown; ask for html in content_format and it also carries content_html. Doc ids are UUID strings, so this takes the id from Shortcut: List Docs or Shortcut: List Epic Docs rather than anything you can read off a URL number.
Shortcut: Get Epic
Fetch one epic in full, including its comment thread, its current health record, and the stats block counting stories and points by state. The stats are what a progress readout wants: num_points_done against num_points, and num_stories_done against num_stories_total.
Shortcut: Get Epic Comment
Fetch one comment on an epic, with its replies nested inside it. Both ids are required, because a comment id is only resolvable in the context of its epic.
Shortcut: Get Epic Health
The CURRENT health status of one epic: On Track, At Risk, Off Track or No Health, with the author and the explanatory note. Health is a deliberate human statement about an epic rather than anything computed from its stories, which is what makes it worth reporting on beside the stats block.
Shortcut: Get Epic Workflow
The workspace's epic workflow and its states. There is exactly ONE epic workflow per workspace, which is why this takes no id and why epic states, unlike story states, are not scoped to a team. Use it to turn an epic_state_id into a name, or to find the id to move an epic to.
Shortcut: Get File
Fetch one uploaded file's metadata: name, size, content type, uploader and the stories it is attached to. This returns the record, not the bytes; fetching the bytes means requesting the url with the same Shortcut token.
Shortcut: Get Iteration
Fetch one iteration with its dates, status, associated teams and stats. The stats block is what a sprint report needs: points done against points total, plus average cycle time and average lead time in SECONDS, which is the unit to convert before charting.
Shortcut: Get Key Result
One key result on an objective: its name, type, progress percentage, and its initial, current and target values. There is no key results lister on this API, so the id comes from the key_result_ids array on Shortcut: Get Objective.
Shortcut: Get Label
Fetch one label with its stats block: how many stories and points carry it, split by state, and how many epics. This is the cheapest way to answer how much work is tagged a given way without listing the stories.
Shortcut: Get Linked File
Fetch one linked file with its URL, type and the stories it is attached to.
Shortcut: Get Member
Fetch one member by UUID: role, state, team memberships and profile. Use this to turn an owner id from a story into a name and an email address for a notification.
Shortcut: Get Objective
Fetch one objective with its state, categories, stats and the ids of its key results. The key result ids are UUIDs and are the only way to reach Shortcut: Get Key Result, which has no lister of its own.
Shortcut: Get Objective Health
The current health status of one objective: On Track, At Risk, Off Track or No Health, with its author and note. Same mechanism as epic health, one level up the hierarchy.
Shortcut: Get Outgoing Webhook
Fetch one registered outgoing webhook. This connector BLANKS the shared secret in the response, on this node and everywhere else, so a credential never reaches an automation's output or its run log. If you need the secret, it is the one you set when creating the integration.
Shortcut: Get Project
Fetch one project with its workflow, abbreviation, followers and stats block.
Shortcut: Get Repository
Fetch one connected repository by id. Note the path parameter is spelled repo, not repository, which is the one place Shortcut abbreviates a resource name in a URL.
Shortcut: Get Story
Fetch one story in full: description, owners, labels, custom field values, estimate, workflow state, and the nested tasks, comments, files, branches, commits and pull requests. This is the only way to read a story's TASKS, which have no lister of their own and arrive in the tasks array here.
Shortcut: Get Story Comment
Fetch one comment on a story with its text, author, reactions and thread position. Both ids are required, because a comment id is only resolvable in the context of its story.
Shortcut: Get Story History
The full audit trail for one story: who changed what and when, as a bare array of history entries. Each entry carries the actor, the timestamp, an actions array whose changes object holds the old and new value of every field touched, and a references array naming the ids that appear in those changes, so a workflow state id can be rendered as a state name without a second call. This is the node behind cycle-time and state-transition reporting: it is the only place the API says when a story entered a given column.
Shortcut: Get Story Link
Fetch one story link: its verb, the two story ids and the subject's current workflow state. There is no story links lister, so the id comes from the story_links array on Shortcut: Get Story.
Shortcut: Get Story Template
Fetch one story template with the exact story contents it applies: name, description, type, workflow state, team, labels, owners and tasks. Read this before creating from a template if you need to know what you are about to inherit.
Shortcut: Get Task on Story
Fetch one checklist task on a story. There is no lister for tasks on this API, so the task id comes from the tasks array on Shortcut: Get Story, or from the response of Shortcut: Create Task on Story.
Shortcut: Get Team
Fetch one team with its member ids, workflow ids, default workflow and story counts. The member_ids array is the authoritative roster, and each id resolves through Shortcut: Get Member.
Shortcut: Get Workflow
Fetch one workflow with its states, their order, their types and how many stories sit in each. This node also powers every workflow-state picker in the connector, which is why those fields ask you to choose a workflow first.
Shortcut: List Categories
Every category in the workspace. Categories group OBJECTIVES, which is the whole of their scope: they are not story labels and they do not apply to epics. Archived categories are included, so filter on the archived flag if you only want live ones.
Shortcut: List Category Objectives
Every objective carrying one category. Shortcut also exposes an identical endpoint under the older milestones spelling; this connector builds only the objectives one, because they return the same records and shipping both would suggest they are different things.
Shortcut: List Custom Fields
Every custom field defined in the workspace, each with the full list of enum values it allows. This is the node to run BEFORE setting a custom field on a story: a story carries field_id plus value_id pairs, and both are UUIDs that only this endpoint can tell you. The field id is the row's id; the value id is the id of one entry in that row's values array.
Shortcut: List Doc Epics
The epics one doc is linked to. The relationship is many to many, so this is the other half of Shortcut: List Epic Docs: one doc can be attached to several epics, and one epic can carry several docs.
Shortcut: List Docs
Every doc the token's member can see, in SLIM form: id, title and app URL only, with no body content at all. That is by design, since doc bodies are large. Fetch the text one at a time with Shortcut: Get Doc. Doc ids are UUID strings, unlike the integer ids most of this API uses.
Shortcut: List Epic Comments
The comment thread on one epic. Epic comments nest differently from story comments: each top-level comment carries its replies in its own comments array, rather than the replies carrying a parent id. Walk the nested array rather than filtering on a parent field.
Shortcut: List Epic Docs
The docs linked to one epic, returned in slim form: id, title and app URL, with no body content. Fetch the text with Shortcut: Get Doc once you have the id.
Shortcut: List Epic Health History
Every health status ever recorded on one epic, newest first, with each one's author, note and timestamp. This is what turns a single current status into a narrative: when the epic first went At Risk, who said so, and what changed.
Shortcut: List Epic Stories
Every story inside one epic, with workflow state, owners, estimate and labels on each. This endpoint returns a bare JSON array with no pagination of any kind, so one call returns everything the token can see. On a large workspace that is a large response. This is also the node behind the story pickers in this connector, which is why several nodes carry an epic field they never send: Shortcut has no unscoped story lister, so a picker has to be narrowed to something.
Shortcut: List Epics
Every epic in the workspace, with its state, owners, teams, labels, deadline, health and a stats block counting the stories and points inside it. This endpoint returns a bare JSON array with no pagination of any kind, so one call returns everything the token can see. On a large workspace that is a large response. Use Shortcut: List Epics (Paginated) instead when the workspace has enough epics for that to matter. This is also the source for every epic picker in this connector.
Shortcut: List Epics (Paginated)
The same epics as Shortcut: List Epics, one page at a time. This is the ONLY page-numbered endpoint on the whole Shortcut API: everything else either returns a bare array or uses the search cursor. The response is an object rather than an array, with the epics at data, a total count, and next holding the next page NUMBER or null on the last page. Ask for page 1 first; there is no page 0.
Shortcut: List Files
Every file uploaded to the workspace, with its name, content type, size, uploader and the stories it is attached to. This endpoint returns a bare JSON array with no pagination of any kind, so one call returns everything the token can see. On a large workspace that is a large response. The url on each row points at Shortcut's own API and needs the same token to fetch, so it is not a link you can hand to someone outside the workspace.
Shortcut: List Iteration Stories
Every story in one iteration, with workflow state, owners, estimate and labels on each. This endpoint returns a bare JSON array with no pagination of any kind, so one call returns everything the token can see. On a large workspace that is a large response. This is the raw material for a standup summary or an end-of-sprint report: group the rows by workflow_state_id and sum the estimates.
Shortcut: List Iterations
Every iteration in the workspace, with its dates, status, teams, labels and a stats block counting points and stories by state, plus average cycle and lead time in seconds. This endpoint returns a bare JSON array with no pagination of any kind, so one call returns everything the token can see. On a large workspace that is a large response. The status field is derived from the dates rather than set by hand: unstarted, started or done.
Shortcut: List Label Epics
Every epic carrying one label. Labels are shared between stories and epics, so the same label id answers both this node and Shortcut: List Label Stories.
Shortcut: List Label Stories
Every story carrying one label, across all teams, epics and iterations. This endpoint returns a bare JSON array with no pagination of any kind, so one call returns everything the token can see. On a large workspace that is a large response. This is the node behind cross-cutting reports that no single board shows, such as everything tagged for a compliance push.
Shortcut: List Labels
Every label in the workspace. By default each row carries a stats block counting the stories, points and epics using that label, which is expensive on a workspace with many labels; set slim to true to drop the stats and get just the label records. This endpoint returns a bare JSON array with no pagination of any kind, so one call returns everything the token can see. On a large workspace that is a large response.
Shortcut: List Linked Files
Every linked file in the workspace, with its URL, integration type and the stories it is attached to. This endpoint returns a bare JSON array with no pagination of any kind, so one call returns everything the token can see. On a large workspace that is a large response. A linked file is a REFERENCE to a document somewhere else, not a copy, so Shortcut holds only the URL and the metadata around it.
Shortcut: List Members
Every member of the workspace, with each one's role, state, teams and profile (name, mention name, email address). Member ids are UUIDs and this is where every owner, follower and author id in the connector comes from. Disabled members are included by default, which matters for a directory sync: filter on the disabled flag rather than assuming the list is current staff.
Shortcut: List Objective Epics
Every epic rolling up into one objective, each with its own state, deadline and stats block. Summing the epics' stats is how you get progress for a whole objective, since the objective's own stats block does not count stories.
Shortcut: List Objective Health History
Every health status ever recorded on one objective, newest first, with author, note and timestamp on each.
Shortcut: List Objectives
Every objective in the workspace, with its state, categories, key result ids and a stats block. This endpoint returns a bare JSON array with no pagination of any kind, so one call returns everything the token can see. On a large workspace that is a large response. Objectives are what Shortcut used to call Milestones: the entity_type on each row still reads milestone, which is a renaming artefact rather than a different kind of record.
Shortcut: List Project Stories
Every story in one project. This endpoint returns a bare JSON array with no pagination of any kind, so one call returns everything the token can see. On a large workspace that is a large response.
Shortcut: List Projects
Every project in the workspace, with its abbreviation, colour, workflow and stats. Projects are Shortcut's older way of grouping stories and most workspaces now organise by TEAM instead, so this often returns an empty array on a newer workspace, which is a correct answer rather than a failure.
Shortcut: List Repositories
Every version control repository connected to the workspace, with its full name, type and URL. This lists what Shortcut's own GitHub, GitLab or Bitbucket integration has connected; it does not connect anything, and the connector cannot add a repository.
Shortcut: List Story Comments
Every comment on one story, including replies (a reply carries the id of its parent in parent_id) and each comment's emoji reactions. Deleted comments are returned with deleted set to true and their text removed, so the length of this array is not a count of live comments.
Shortcut: List Story Sub-tasks
The stories linked as sub-tasks of one parent story, returned as full story records rather than as ids. Sub-tasks are STORIES in Shortcut, which is the thing to know before wiring this up: they have their own workflow state, owners and estimate. That makes them different from a story's TASKS, which are checklist items living inside the story and readable through Shortcut: Get Story. Sub-tasks require the workspace feature of the same name.
Shortcut: List Story Templates
Every story template in the workspace, each with its author, last-used timestamp and the full story contents it will apply. Feed a template id from here into Shortcut: Create Story From Template. If this returns an error rather than an empty list, the workspace has story templates switched off.
Shortcut: List Team Stories
Every story owned by one team. This is the ONLY endpoint on the Shortcut API that takes limit and offset, and it still returns a bare array with no total, so there is no way to know you have reached the end except by receiving fewer rows than you asked for. Leave both blank to get everything in one call.
Shortcut: List Teams
Every team in the workspace, with its members, workflows, mention name, colour and story counts. Teams are called Groups throughout this API and their ids are UUIDs, which is what to check first when a group_id is rejected: an integer there is always the wrong kind of id. This is the source for every team picker in the connector.
Shortcut: List Workflows
Every story workflow in the workspace, each with its full list of states in board order. Run this first when a workflow_state_id is rejected: states belong to ONE workflow, a team works in a specific set of workflows, and a state id borrowed from another team's board is not valid. Each state carries a type of unstarted, started or done, which is the portable way to ask what a column means across teams that named theirs differently.
Shortcut: Search Docs
Find docs by TITLE. This is the odd one out among the search endpoints: it takes a title rather than a query, understands none of Shortcut's search operators, and searches titles only rather than doc bodies. It adds three filters the others do not have: archived, created by me and followed by me, where me is the member who owns the API token.
Shortcut: Search Epics
Find epics with a Shortcut search query. Same operators and same cursor contract as Shortcut: Search Stories, restricted to epics.
Shortcut: Search Everything
Run one Shortcut search query across several entity types at once. The response is an OBJECT with a separate result envelope per type (stories, epics, iterations and objectives), each with its own total, data array and next value, rather than one merged list. Defaults to stories and epics; set entity_types to widen or narrow it. Result ordering can shift between calls as search ranking decays and new work is created, which is exactly why paging must follow the next value rather than counting pages.
Shortcut: Search Iterations
Find iterations with a Shortcut search query. Worth knowing that Shortcut: List Iterations returns every iteration in one call with no query at all, so this node earns its place only when you want the search operators, for instance to find iterations belonging to one team.
Shortcut: Search Objectives
Find objectives with a Shortcut search query. Shortcut also exposes the identical search under the older milestones spelling; this connector builds only the objectives one, because both return the same records.
Shortcut: Search Stories
Find stories with a Shortcut search query, the same syntax the web app's search box takes: owner:example, state:"In Progress", is:blocked, label:security, created:2026-01-01..2026-02-01, and so on. Use this when a human would type the question; use Shortcut: Search Stories by Field when a machine is assembling the filters. The response carries total, data and next.
Shortcut: Search Stories by Field
Find stories by structured field values rather than by a text query: every story owned by a member, in an epic, in a workflow state, created between two dates. Every filter is optional and the ones you set are combined, so this is the node for report-style questions that Shortcut: Search Stories cannot express in a query string. It is a POST because the filter set is a body rather than a URL, but it reads and changes nothing. The response is a bare array with no pagination and no total, so a broad filter returns a broad response.
Action nodes 60
Shortcut: API Request
Call any Shortcut v3 endpoint directly, with the connector's credential applied for you. This exists so nothing on the API is unreachable: the endpoints this connector does not ship as their own node, and anything Shortcut adds after it was built. Give it a method and a path such as /api/v3/epics, plus a query object or a JSON body if the endpoint takes one. It is marked Safe Read because a GET is, but the method is yours to choose: a DELETE typed here deletes, with none of the warnings the dedicated nodes carry. Prefer a dedicated node whenever one exists, because those bring validated fields, pickers and sample outputs.
Shortcut: Add Reaction to Story Comment
Add an emoji reaction to a story comment, as the member who owns the API token. The response is the comment's full reaction list after the change rather than just the one added. Reacting twice with the same emoji is not an error.
Shortcut: Create Category
Create a category for objectives. Only the name is required. Note you rarely need this node: attaching a category by name through Shortcut: Create Objective or Update Objective creates it on the spot if it does not exist. Use this one when you want to set the colour up front, or to pre-create the vocabulary before anything uses it.
Shortcut: Create Doc
Create a doc. Title and content are both required. Storage is always HTML: content_format tells Shortcut what you are SENDING, and markdown is converted on the way in. The default is html, which is the trap on this endpoint, because Markdown posted without setting content_format is stored as literal text with its hashes and asterisks showing. Set markdown explicitly when you are sending Markdown.
Shortcut: Create Epic
Create an epic. Only the name is required. Two pairs of fields on this endpoint are legacy and this node deliberately exposes only the current one of each: use epic_state_id rather than state, because Shortcut ignores state whenever epic_state_id is set, and use objective_ids rather than milestone_id, because Milestones were renamed Objectives and the old field is marked deprecated in Shortcut's own API description. Stories are added to an epic from the story side, by setting epic_id on Shortcut: Create Story or Update Story.
Shortcut: Create Epic Comment
Post a top-level comment on an epic. To reply to an existing comment instead, use Shortcut: Reply to Epic Comment, which takes the parent comment's id in the path. As with story comments, the comment is attributed to the token owner unless author_id names another member.
Shortcut: Create Iteration
Create an iteration. Name, start date and end date are all required, and the dates are plain calendar dates (2026-02-01), not timestamps. This is the node to schedule if you want sprints created ahead of time rather than by hand. Stories are pulled in afterwards by setting iteration_id on Shortcut: Update Story or Shortcut: Update Stories in Bulk.
Shortcut: Create Label
Create a label. Only the name is required. Labels are created implicitly whenever a story or epic is given a label name that does not exist yet, so use this node when you want to set the colour and description up front, or to seed a vocabulary before anything writes to it. Creating a label that already exists by name is rejected.
Shortcut: Create Linked File
Attach an external document to Shortcut by URL. Name, type and url are all required. This is the node that gets a spec, a design file or a dashboard onto a story without uploading anything: set story_id and the link appears on that story. Use type url for a plain web link and the named integration types when the document really does live in that service, since the type only affects how Shortcut labels and icons the link.
Shortcut: Create Objective
Create an objective, the strategic goal that epics roll up into. Only the name is required. Epics are attached from the epic side, by setting objective_ids on Shortcut: Create Epic or Update Epic, so a fresh objective starts empty.
Shortcut: Create Outgoing Webhook
Register a URL for Shortcut to POST its activity events to. This is the OUTBOUND direction: Shortcut calls you when something changes, which is how a workflow reacts to Shortcut in real time instead of polling. Set the secret and verify the signature on every delivery, because the endpoint is otherwise open to anyone who learns the URL. Note the response is 204 with no body, so the integration's own id is not returned here: list or read it afterwards to get it.
Shortcut: Create Project
Create a project. Name and team_id are both required. Note team_id here is the INTEGER id Shortcut's API uses for a workflow grouping, not the UUID that identifies a Team in the modern sense: read it from the team_id field on Shortcut: List Workflows. If your workspace organises by team rather than by project, you probably want Shortcut: Create Team instead.
Shortcut: Create Stories in Bulk
Create many stories in one request. Each element takes exactly the same fields as Shortcut: Create Story, so this is the node to use when an upstream step produced a list rather than a single item. The response is the array of created stories in the order they were sent. There is no documented partial-success mode: a malformed element fails the whole request, so validate the array upstream rather than relying on the good elements landing.
Shortcut: Create Story
Create one story: a feature, a bug or a chore. Only the name is required, so a minimal call files an unestimated story into the workspace's default workflow state. Everything else is optional and can be set later with Shortcut: Update Story. Two fields are worth setting on creation rather than after: workflow_state_id, because a story created without one lands in the default state rather than where you meant it, and group_id, because the team owning a story governs which workflow its states come from. The response is the created story, so the new id is available to the rest of the automation immediately.
Shortcut: Create Story Comment
Post a comment on a story, or reply to an existing comment by setting parent_id. The comment is attributed to the member who minted the API token unless author_id names someone else, which matters when a shared automation posts on a team's behalf: without author_id every comment appears to come from one person.
Shortcut: Create Story From Template
Create a story from a story template, applying the template's name, description, tasks, labels, owners and workflow state, then layering your own changes on top. This is the node that actually uses a template; setting story_template_id on Shortcut: Create Story only records an association and copies nothing. The fields here are the ADD and REMOVE variants on purpose: Shortcut's own documentation says the add and remove forms cannot be combined with their whole-value counterparts on the same request, so this node exposes the additive forms and leaves wholesale replacement to Shortcut: Update Story afterwards.
Shortcut: Create Story Link
Relate two stories. All three fields are required and the DIRECTION matters: the sentence reads subject verb object, so blocks with subject 100 and object 200 means story 100 blocks story 200, and reversing the ids reverses the meaning. Shortcut shows the inverse on the other story automatically, so there is no need to create the mirror link yourself.
Shortcut: Create Task on Story
Add a checklist task to a story. A task is a line item inside the story rather than a story of its own: it has a description, a complete flag and optional owners, and it never has a workflow state or an estimate. Use sub-tasks instead when the work needs to move across a board on its own.
Shortcut: Create Team
Create a team. Name and mention_name are both required, and the mention name is what people type after @@ in comments, so it should be short and unique. Give the team at least one workflow id: a team with no workflow has no board for its stories to live on, and Shortcut will not infer one.
Shortcut: Delete Category
Permanently delete one category. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. It is removed from every objective carrying it, and those objectives are otherwise untouched. Archiving through Shortcut: Update Category is almost always the better move, because it retires the category without rewriting history.
Shortcut: Delete Doc
Permanently delete one doc and everything attached to it. Shortcut's own wording is permanent, and it requires ADMIN access to that doc, so a token whose member can read and edit the doc may still be refused here. There is no archive state for docs, so read the content out with Shortcut: Get Doc first if you might want it back.
Shortcut: Delete Epic
Permanently delete one epic, with its comment thread and health history. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. If you want it reversible, set archived to true through the matching Update node instead. The epic's STORIES are not deleted: they survive with their epic_id cleared, which means they leave the epic's board silently rather than disappearing. Nothing in the API tells you afterwards which stories used to belong to it, so list them first if you need that.
Shortcut: Delete Epic Comment
Delete one comment from an epic. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. Deleting a comment that has replies leaves the thread in place with the parent marked deleted and its text gone.
Shortcut: Delete File
Permanently delete one uploaded file. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. The bytes go with it, so this is the least recoverable delete in the connector: a story or comment can be recreated from an export, an image cannot. It is removed from every story it was attached to, and this connector cannot re-upload it, because Shortcut's upload endpoint is a multipart form a canvas has no file for.
Shortcut: Delete Iteration
Permanently delete one iteration. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. An iteration has no archived state on this API, so there is no reversible alternative to reach for. The STORIES in it are not deleted: they survive with their iteration_id cleared, which also erases the record of which sprint they were worked in, so export the sprint's story list before deleting if the history matters.
Shortcut: Delete Label
Permanently delete one label. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. It is removed from EVERY story and epic carrying it, and nothing records that it was ever there, so a saved view or a report filtered on that label silently returns nothing afterwards. Archiving through Shortcut: Update Label retires it without that cost and is almost always the right choice.
Shortcut: Delete Linked File
Delete one linked file. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. The DOCUMENT ITSELF is untouched, because Shortcut only ever held a URL, so this is the most recoverable delete in the connector: relink it with Shortcut: Create Linked File if the URL is still known.
Shortcut: Delete Objective
Permanently delete one objective, with its key results and health history. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. If you want it reversible, set archived to true through the matching Update node instead. The objective's EPICS are not deleted: they survive with this objective removed from their objective_ids, so a quarter's worth of work quietly loses its strategic parent rather than disappearing.
Shortcut: Delete Outgoing Webhook
Unregister an outgoing webhook. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. Deliveries stop immediately and silently: nothing downstream is told, so anything listening simply stops hearing from Shortcut. Recreate it with Shortcut: Create Outgoing Webhook, which needs the URL and the secret again.
Shortcut: Delete Project
Permanently delete one project. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. If you want it reversible, set archived to true through the matching Update node instead. Shortcut refuses to delete a project that still contains stories, so the usual sequence is to move or delete the stories first; that refusal is a useful guard rather than an obstacle, since it is the only thing standing between a misfired automation and a lost board.
Shortcut: Delete Stories in Bulk
Permanently delete several stories at once. Shortcut only deletes stories that are ALREADY ARCHIVED here, which makes this the one delete on the whole API with a built-in guard: a live story is not removed by accident. The intended sequence is archive with Shortcut: Update Stories in Bulk, review, then delete. Comments, tasks and history go with the stories and there is no undelete.
Shortcut: Delete Story
Permanently delete one story, with its comments, tasks and history. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. If you want it reversible, set archived to true through the matching Update node instead. Deleting a story that other stories link to also removes those links, and deleting a parent story does not delete its sub-tasks: they are left unparented.
Shortcut: Delete Story Comment
Delete one comment from a story. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. Deleting a comment that has replies leaves the thread in place with the parent marked deleted and its text gone, so the replies survive but their context does not.
Shortcut: Delete Story Link
Remove a relationship between two stories. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. Both STORIES are untouched: only the link between them goes, and it can be recreated with Shortcut: Create Story Link as long as you still know the verb and the two ids.
Shortcut: Delete Task on Story
Delete one checklist task from a story. This is a real delete, not an archive. Shortcut answers 204 No Content and exposes no undelete. A task has no archived state, so unlike a story there is no reversible alternative. Read it first if the text matters.
Shortcut: Link Doc to Epic
Attach a doc to an epic. Both ids go in the URL and there is no body: the relationship IS the request. Linking a doc that is already linked is not an error, so this is safe to run repeatedly, which matters when a retried automation would otherwise duplicate the link.
Shortcut: Remove Reaction from Story Comment
Remove the token owner's emoji reaction from a story comment. Listed as reversible rather than destructive because re-adding the same reaction restores the state exactly. It removes only the token owner's reaction, never another member's.
Shortcut: Reply to Epic Comment
Reply to an existing epic comment, threading the new comment under it. Note the shape of this endpoint: it POSTs to the PARENT COMMENT's own URL, which is the same URL the update node PUTs to and the delete node DELETEs. Same address, three different meanings depending on the method.
Shortcut: Set Epic Health
Record a new health status on an epic. This APPENDS to the epic's health history rather than overwriting the last entry, so the trail of how a piece of work went is preserved. Set No Health to clear the indicator without deleting the history.
Shortcut: Set Objective Health
Record a new health status on an objective. This APPENDS to the objective's health history rather than overwriting the last entry, so the record of how a goal tracked over a quarter stays intact.
Shortcut: Unlink Doc from Epic
Detach a doc from an epic. This removes the LINK only: the doc itself is untouched and still exists in the workspace, which is why this is a reversible write rather than a destructive one. Re-link it any time with Shortcut: Link Doc to Epic.
Shortcut: Update Category
Rename, recolour or archive one category. Archiving keeps it on the objectives already carrying it while removing it from the pickers offered for new ones, which is the safe way to retire a category that is no longer used.
Shortcut: Update Custom Field
Change a custom field definition: rename it, enable or disable it for the workspace, reorder it, or edit the list of values it allows. Treat the values field carefully. It is the FINAL STATE of the value list, so read the field first with Shortcut: Get Custom Field and send back the existing entries with your additions, or you will drop the ones you left out. Disabling a field hides it from new stories without touching the values already recorded on existing ones, which is the reversible way to retire one.
Shortcut: Update Doc
Change a doc's title, its body, or both. Content REPLACES the whole document rather than appending, and there is no revision endpoint on this API to recover the previous text from, so read the doc first if you are rewriting rather than generating. content_format carries the same default-to-html trap as Create Doc: Shortcut's own documentation flags it as a warning.
Shortcut: Update Epic
Change one epic: move it through the epic workflow, retitle it, reassign it, archive it. Only the fields you fill in are sent. The list fields REPLACE rather than append, so sending one owner id leaves the epic with exactly one owner. Moving an epic through the workflow does not move its stories: an epic marked done can still hold unfinished stories, and Shortcut allows that deliberately.
Shortcut: Update Epic Comment
Rewrite the text of one epic comment. Text is the only field this endpoint accepts, and the previous text is not recoverable through the API.
Shortcut: Update File
Change an uploaded file's metadata: rename it, describe it, reattribute it or stamp it with an external id. The bytes are immutable, so this node cannot replace a file's contents.
Shortcut: Update Health Record
Correct an existing health record in place, rather than appending a new one. Use this to fix a typo in a note or a status set by mistake; use Shortcut: Set Epic Health or Shortcut: Set Objective Health when the situation has genuinely changed, because those preserve the trail and this overwrites it. One endpoint serves both epic and objective health, so the id decides which one you are editing.
Shortcut: Update Iteration
Rename, re-date or re-scope one iteration. Only the fields you fill in are sent. Moving the dates changes the iteration's derived status, so an iteration can go from started back to unstarted by pushing its start date forward. Stories already in the iteration stay in it whatever the dates do.
Shortcut: Update Key Result
Update a key result's name or its measured values. This is the node that makes an objective self-updating: point a scheduled automation at your metric source and push the observed value in, and the progress percentage recalculates from the initial, observed and target values. Each value is an object rather than a bare number, because a key result may be numeric, a percentage or boolean.
Shortcut: Update Label
Rename, redescribe, recolour or archive one label. Renaming updates it everywhere at once, since a label is a single record that stories and epics point at rather than a string copied onto each. Archiving hides it from pickers while leaving it on the work already carrying it.
Shortcut: Update Linked File
Change a linked file: point it at a new URL, rename it, or move it to a different story. Repointing the URL is the useful case, because a document that moved does not have to be relinked everywhere it appears.
Shortcut: Update Objective
Change one objective: rename it, move its state, re-categorise it, archive it, or reorder it against its siblings. Only the fields you fill in are sent. The categories list REPLACES the current set rather than adding to it.
Shortcut: Update Project
Rename, recolour, re-home or archive one project. Only the fields you fill in are sent. Archiving a project is the reversible way to retire it and leaves its stories in place.
Shortcut: Update Stories in Bulk
Apply one change to many stories at once: move a batch to a workflow state, pull a set into an iteration, add an owner across a list. Only story_ids is required and every other field is optional, so send just the ones you mean to change. This node uses ADD and REMOVE lists rather than whole-value replacement for owners, followers, labels and custom fields, which is the important difference from Shortcut: Update Story: adding an owner here does not wipe the owners a story already had.
Shortcut: Update Story
Change one story: move it between workflow states, reassign it, re-estimate it, retitle it, archive it. Only the fields you fill in are sent, so an unfilled field is left alone. The list fields here REPLACE rather than append, so sending one owner id leaves that story with exactly one owner. Use Shortcut: Update Stories in Bulk when you want add and remove semantics, even for a single story. The sub_tasks field is the same trap in a sharper form: Shortcut treats it as the final state of the parent's sub-tasks, so anything missing from the array is unlinked.
Shortcut: Update Story Comment
Rewrite the text of one story comment. Text is the only field this endpoint accepts: author, timestamps and thread position are fixed once a comment exists. The previous text is not recoverable through the API, so read it first if you need it.
Shortcut: Update Story Link
Change an existing story link: swap the verb, or repoint either end. Swapping subject and object is how you reverse a relationship that was created the wrong way round, which is the common reason to reach for this node.
Shortcut: Update Task on Story
Tick, retitle, reassign or reorder one checklist task on a story. Ticking is the common use: set complete to true. The owner_ids list replaces the task's current owners rather than adding to them.
Shortcut: Update Team
Change one team: rename it, change its membership or workflows, or archive it. The member_ids and workflow_ids lists REPLACE the current sets rather than adding to them, so read the team first and send the full intended roster: sending one member id leaves the team with exactly one member. Archiving is the only way to retire a team, because this API has no delete for one.
Connecting Shortcut takes one credential
Generate a token in Shortcut under Settings, Account, API Tokens. There is no OAuth flow and no app to register: the token is a single value you paste into Mattering once. The decision worth making before you generate it is whose token it should be, because a Shortcut token carries the permissions of the member who minted it and nothing more. Anything that member cannot see in the web app is invisible to every node, and comments written through the connector are attributed to them unless a node names another author, so a shared automation posting under one person's name is a choice rather than an accident. Shortcut's Free plan is perpetual for workspaces of ten users or fewer, and its pricing table lists API and Webhook Access as included on all four plans, so a working credential does not require a paid subscription.
Decide which Shortcut member holds the connection. The token inherits that member's own access and attribution, and unattended workflows keep running on it, so a member the team controls beats an individual's personal account.
In Shortcut, open Settings, then Account, then API Tokens, and generate a token. Copy it straight away and treat it as a password: it grants everything that member can do.
In Mattering Automations, open Connectors, click Add Connector and choose Shortcut, then Create new credential and paste the token into the API token field. It is stored encrypted and never appears in workflow logs.
Click Test Connection. Mattering calls the member endpoint, the cheapest authenticated call on the API, and reports back which member the token acts as. A 401 means the token was revoked, the member was deactivated, or the value was pasted with whitespace around it, since Shortcut tokens do not expire on their own. A 403 means the token is valid and that member's role forbids the call. A 429 means the 200 per minute limit, so wait a minute.
All 126 Shortcut nodes appear in the canvas palette and over MCP. Run Shortcut: Get Current Member first: it confirms the identity behind the credential and it is also the only place the API publishes your workspace's estimate scale, which a story-creating workflow needs before it sets an estimate.
Before your first write, run Shortcut: List Workflows and Shortcut: Get Workflow for the team you are targeting. Workflow states are scoped to one workflow in Shortcut, so a state id borrowed from another team's board is rejected, which is why the state pickers in this connector ask you to choose a workflow first.
Common uses for the Shortcut MCP
A sprint review that assembles itself before the meeting
Shortcut: List Iterations returns every iteration with its dates, status and a stats block counting points and stories by state, and Shortcut: List Iteration Stories pulls the work inside the one you care about. An AI step can then turn that into a plain summary of what shipped, what slipped and what is still unstarted, and post it to Slack or open a Mattering Productivity task for the person running the meeting. Everything in this path reads, so it is safe to schedule and safe to hand to an assistant. Note that an iteration's status is derived from its dates rather than set by hand, so unstarted, started and done mean exactly what the calendar says.
Turning inbound reports into properly filed stories
A form submission, a support ticket or an alert can become a real Shortcut story without anybody retyping it. Shortcut: Create Story needs only a name, but two fields are worth setting at creation rather than after: the workflow state, because a story created without one lands in the default state rather than where you meant it, and the team, because the team owning a story governs which workflow its states come from. Shortcut: Create Task on Story lays down the triage checklist, Shortcut: Create Story Link relates it to the story it blocks or duplicates, and Shortcut: Create Stories in Bulk handles the case where the upstream step produced a list. If your team files from a template, use Shortcut: Create Story From Template, because setting a template id on an ordinary create only records the association and copies nothing.
Keeping Shortcut and another system pointing at each other
Shortcut: Find Stories by External Link is the node that makes a two-way sync tractable: given a URL in another system it answers which Shortcut stories point at it, with no search query to construct. Pair it with Shortcut: Update Story to push status changes back, and with Shortcut: Create Linked File to attach a spec, a design file or a dashboard to a story by URL without uploading anything. The match is on the whole URL, so normalise trailing slashes and query strings on your side before you rely on it. This is where chaining pays: the same workflow can read the other system through its own connector, so the sync is one canvas rather than two integrations that have to agree.
Cycle-time reporting that Shortcut's own reads cannot assemble
Shortcut: Get Story History is the only place this API says when a story entered a given column: it returns the full audit trail, with the actor, the timestamp, the old and new value of every field touched and a references array so a workflow state id can be rendered as a state name without a second call. Shortcut: Search Stories by Field is how you pick the population, because it filters by structured values rather than a query string, including a created-between window, which makes it the one endpoint on this API a scheduled incremental read can drive. The catch worth designing around is that history is per story, so a series is assembled by fanning out, and at 200 requests per minute a fan-out over a whole quarter needs pacing rather than a tight loop.
Objectives that update themselves from the numbers
Shortcut: Update Key Result pushes an observed value into a key result and Shortcut's progress percentage recalculates from the initial, observed and target values, so a scheduled workflow that reads your metric source keeps the objective honest without anybody editing it by hand. Shortcut: Set Objective Health and Shortcut: Set Epic Health append a status and a note to the health trail, which is the right node when the situation genuinely changed, while Shortcut: Update Health Record corrects an existing entry in place and overwrites the trail, which is the right node only for a typo. One planning note before you build on this: Shortcut's pricing table lists Strategic Objectives and Key Results as included on the Business and Enterprise plans, while Tactical Objectives are on every plan, so check your plan before the key result half of this becomes load bearing.
Backlog hygiene with a guard rail in front of it
The safe pattern on this API is archive, review, then delete, and Shortcut built the guard for you. Shortcut: Update Stories in Bulk archives a set of stale stories with add and remove semantics rather than the whole-value replacement that the single-story update uses. A human or a scheduled digest reviews what was archived. Only then does Shortcut: Delete Stories in Bulk remove them, and it refuses anything that is not already archived, which makes it the one delete on the whole API that cannot take a live story by accident. Apply the same instinct to labels and categories: Shortcut: Update Label with archived set to true retires a label without stripping it from every story and epic carrying it, which is what Shortcut: Delete Label does and does not record.
Example: A Monday iteration digest, with the cleanup staged behind a person
Built on the Automations canvas in a few minutes, or described in plain language to Matter and built for you.
Trigger: a scheduled node fires every Monday morning. This connector ships no Shortcut trigger node, so workflows here run on a schedule, or on Mattering's generic Incoming Webhook trigger if you register a URL with Shortcut: Create Outgoing Webhook and verify the signature yourself.
Shortcut: List Iterations returns every iteration with its stats block, and a condition picks the one whose status is started.
Shortcut: List Iteration Stories reads the work inside it in a single call, which on this API returns the whole collection rather than a first page.
Shortcut: Search Stories by Field runs alongside it to pull anything created in the last week that is still unstarted and unowned, filtered by structured values rather than a query string.
An AI step turns both lists into a short readout: what moved, what is blocked, what has no owner. Nothing has been written to Shortcut yet.
The readout is posted to the team channel and opened as a Mattering Productivity task for whoever runs the review, so a person decides what to do about it.
On approval, Shortcut: Update Stories in Bulk archives the stories the team agreed are dead. Deletion is a separate, later decision, because Shortcut: Delete Stories in Bulk is permanent and only ever runs against work that is already archived.
Shortcut MCP questions, answered
Does Shortcut publish its own MCP server, and should I use that instead?
Yes, and for some jobs you should. Shortcut hosts a remote MCP server at mcp.shortcut.com, and it authenticates with OAuth rather than an API token, so your client prompts you to sign in to Shortcut on first use. Both of Shortcut's MCP repositories on GitHub are now archived, with the README stating that all future development moved to the remote server, and Shortcut's pricing table lists MCP Server as included on every plan including Free. If what you want is a Shortcut-aware assistant inside your editor, install theirs and be done. This connector answers a different question. The same 126 nodes are steps on an automation canvas, so a workflow runs on a schedule or an inbound event with nobody present, chains Shortcut with every other connector in one place rather than making Shortcut the only tool in the conversation, holds one credential in a shared workspace store instead of in each person's local config, and marks the 16 nodes that destroy something so an assistant or a reviewer can be held back from them. Those are the honest differences; the vendor server is a good tool and this page is not going to pretend otherwise.
Does this work with Claude, ChatGPT, or other AI assistants?
Yes. Mattering exposes these nodes over MCP, the open protocol for connecting AI assistants to tools, so any MCP-capable assistant, including Claude and ChatGPT, can connect to your workspace and work with Shortcut through them. Matter, the assistant built into the platform, uses the same nodes natively. Because the nodes are also canvas steps, the workflow an assistant helps you design can then run unattended. The split worth making is by category: the 66 read nodes are low risk to hand an assistant, while the writes, and especially the 18 that delete, are better behind a condition or an approval step, since Shortcut exposes no undelete for anything.
What can this connector actually do, and what is not included?
126 nodes: 66 read and 60 act, built one for one from 125 of Shortcut's 143 v3 operations, plus one catch-all. Reads cover stories, sub-tasks, tasks, comments, story history and story links, epics with their comments, docs and health history, iterations, objectives, key results, docs, files, linked files, labels, categories, projects, teams, workflows, custom fields, members, repositories and all six search endpoints. Writes cover creating and updating stories singly and in bulk, tasks, comments and replies, reactions, story links, epics, iterations, objectives, docs, teams, projects, labels, categories, linked files, health records and outgoing webhooks, plus 18 deletes. Eighteen operations are deliberately not built. Eight are milestone endpoints, six of which Shortcut's own document marks deprecated with the objective equivalent named, because Milestones and Objectives are the same entity under two names. File upload is out because it is a multipart form and a canvas has no local file to attach. The four workspace toggles that enable or disable Iterations and Story Templates for every member are out, because that is an administrator's act in the web app. Deleting a custom field definition is out because it wipes that field's values from every story in the workspace with no undelete, so Shortcut: Get Custom Field and Shortcut: Update Custom Field are there and the delete is not. Story template creates, updates and deletes are out because a template body is a nested document authored in Shortcut's own editor. Nothing is unreachable regardless: Shortcut: API Request is an advanced node that calls any Shortcut v3 endpoint directly.
Do I need a paid Shortcut plan?
No. Shortcut's pricing table lists API and Webhook Access as included on all four plans, Free, Team, Business and Enterprise, and the Free plan is perpetual rather than a trial: Shortcut states that if you have ten or fewer users when the trial ends you can simply keep using Shortcut for free. A few limits in that table do shape what some nodes can reach, and it is better to read them now than to debug them later. The Free plan is capped at one team and one customizable workflow, and the Team plan at five of each, which matters for Shortcut: Create Team and for workflows generally. Strategic Objectives and Key Results are listed on Business and Enterprise only, while Tactical Objectives appear on every plan, so treat Shortcut: Get Key Result and Shortcut: Update Key Result as plan dependent. Advanced Custom Fields are Business and Enterprise, while ordinary custom fields and labels are on every plan.
What are Shortcut's rate limits?
Shortcut documents 200 requests per minute and answers HTTP 429 past that, and its documentation specifies no Retry-After header on the response. The missing header is the part that catches people: a retry loop written against the usual contract reads nothing, treats it as zero and hits the limit again immediately, so back-off against this API must be fixed or exponential rather than header driven. This connector surfaces the 429 as an ordinary unsuccessful result with that fact attached rather than retrying inside the node, so a workflow can branch on it instead of a hidden retry burning the whole minute. The structural fix is to prefer the collection listers, since most of them return everything in one response, over a loop of single reads.
Why does one list node return everything and another return a page?
Because this API has four pagination shapes and the connector declares which one each node uses rather than guessing. 35 of the 70 read endpoints return a bare JSON array with no envelope, no page parameter and no total, so Shortcut: List Epics and Shortcut: List Labels return the whole collection in one response. That is fine for most workspaces and a large payload for a big one, which is exactly why Shortcut: List Epics (Paginated) exists for epics alone and is the only page-numbered endpoint on the API. The six search nodes use a cursor that is not an opaque token but the path and query string of the next page, because result ordering shifts as ranking decays and new work is created, and they are capped at 1000 total matches, past which Shortcut returns a dedicated error rather than an empty page, so narrow the query rather than paging further. And Shortcut: List Team Stories is the only endpoint that takes limit and offset, over a bare array with no total, so you know you have finished only by receiving fewer rows than you asked for.
Can a workflow react to changes in Shortcut in real time?
Not through a Shortcut trigger node, because this connector does not ship one. Shortcut workflows here run on a schedule, which covers most reporting and hygiene jobs well. If you need to react to an event, Shortcut: Create Outgoing Webhook registers a URL for Shortcut to post its activity events to, and you can point that at Mattering's generic Incoming Webhook trigger. Two things about doing it that way are yours rather than ours: set the secret on the registration and verify the signature on every delivery, because a webhook URL is otherwise open to anyone who learns it, and note that the registration call answers 204 with no body, so read the integration back afterwards if you need its id for Shortcut: Delete Outgoing Webhook later.
What happens when a workflow deletes something in Shortcut?
It is gone. Every delete on this API answers 204 No Content and Shortcut documents no undelete for any resource, so 16 of the 18 delete nodes are flagged destructive and each one states its own blast radius rather than hiding behind a general warning. Deleting an epic does not delete its stories, but they lose their epic id and nothing afterwards records which ones belonged to it, so list them first. Deleting an objective leaves its epics but removes the objective from them. Deleting a label strips it from every story and epic carrying it, and a saved view filtered on that label then silently returns nothing. Deleting an uploaded file takes the bytes with it and this connector cannot re-upload them, because Shortcut's upload endpoint is a multipart form. Deleting a doc requires admin access to that doc and Shortcut calls it permanent. Two are gentle enough to be flagged as reversible writes instead: unlinking a doc from an epic leaves the doc untouched, and removing a reaction removes one emoji. Where archiving exists, the delete node points at it, because setting archived to true through the matching update node is almost always what somebody actually wanted.
Put Shortcut to work inside your workspace
Connect once, then use every node in visual automations and AI agent workflows across the Mattering platform.
All product names, logos, and brands are property of their respective owners and are used for identification purposes only. Use of these names, logos, and brands does not imply endorsement or affiliation. Shortcut is a product of its respective owner; Mattering is an independent platform.