Claude Port IO connectorWhat Claude can do in your Port portal.
The Claude Port IO connector exposes 12 tools: 7 read your catalog, 5 change it, 0 are undocumented. Here: what Claude can look up in Port, what it can run or edit, what it asks first, what your Port role decides, and who can switch it off.
Verified Trustpilot reviews · AI, automation & growth agency
What changes when Claude can query your Port catalog
Finding who owns a service, which scorecard it fails or which action deploys it usually means clicking through Port's catalog pages. With this connector you ask Claude in plain words: it queries blueprints, entities and scorecards itself, and, if your Port role allows it, runs a self-service action or edits the catalog model. Port decides which of those tools you even see, based on your permissions there.
Ownership answered in one question. You need the owner of a service: list_entities queries the services blueprint with filters, sorting and grouping, without anyone opening the catalog.
Standards checked before a release. You want to know what a scorecard checks: list_scorecards returns its complete rule configuration.
A governed action, launched from the chat. You need a self-service action: list_actions shows the self-service actions available with their inputs, then run_action runs the one that fits.
Now what the directory sheet does not say. Port splits its tools by role: upsert_blueprint, upsert_scorecard and upsert_action belong to the Builder set meant for platform engineers and admins, not to the Developer set. Port documents ways to hide all write tools or restrict which actions can run, but describes them for token-based setups, not for this directory connector. And nothing runs on its own: no tool wakes up when a scorecard drops. For background reactions you need an automation tool, a different job; the Integrations hub gathers that side.
The vocabulary in one minute
Five words you will meet while connecting Port.
- Connector
- The link you set up once between Claude and an account you already own, so Claude can work in it while it answers you.
- Tool
- One named thing a connector lets Claude do in that account. Claude picks the ones it needs; the directory sheet lists them by name.
- Authorization
- Port's own sign-in screen, where you hand Claude the access it will use. Granted once per person, and it can be taken back.
- Approval
- The confirmation Claude waits for before it goes ahead with something that changes your portal, shown in the chat when it matters.
- MCP
- The shared standard connectors are built on: it is what lets an assistant such as Claude talk to a service outside it.
Setting up the Claude Port IO connector in three steps
- 01
Find Port IO in Claude
Open Claude's settings, go to Customize, then Connectors, and look for Port IO in the list. On a Team or Enterprise workspace, an Owner or Primary Owner enables the connector before each member can sign in on their own.
- 02
Connect from its row
Click Connect on the Port IO row, then sign in to Port in the window the service opens itself. If the link breaks later, Disconnect and connect again.
- 03
Read the authorization screen
Read the authorization screen before you accept. It belongs to Port, not to Claude, and it is what decides how far the access reaches. Port adds that the tools you then see depend on your permissions in Port.
The 12 tools, split by what they do
Port IO gives Claude 12 tools: 7 that read your account, 5 that change something in it.
Two groups: what Claude reads in your portal, and what it runs or changes there. Names stay exactly as Claude displays them.
- 7 read
- 5 write
What Claude reads (7)
7 toolsSeven tools that query the catalog, the docs or Port's guidance and leave everything as it is.
list_blueprints
Lists the blueprints of your organization as a summary or, for the ones you name, with property definitions, schemas and allowed values.
list_entities
Queries the entities of a blueprint with filters, sorting and pagination, and can also group values or return only a total count.
list_scorecards
Lists your scorecards in summary, or with their complete rule configuration for the ones you ask about. The directory sheet lists scorecard analysis, next to service health, among the connector's uses.
list_actions
Lists the actions and automations of your organization, with their full input schemas when you name specific ones. It is the reading side of run_action, which then runs one of them by its identifier.
search_port_knowledge_sources
Searches the official Port documentation and returns the most relevant sections with their links and content, so Claude can answer questions about Port itself from its official documentation.
load_skill
Loads Port's specialised guidance for a given task: step-by-step instructions for recurring work such as troubleshooting an integration. It loads guidance and leaves your portal as it is.
list_integrations
Lists the integrations installed in your organization, and for the ones you name, their complete configuration and mapping.
What Claude changes (5)
5 toolsFive tools that run actions or alter the catalog model. No source spells out their confirmation, so the general rule below applies.
run_action
Approval: see the ruleRuns a self-service action or automation by its identifier. The directory sheet describes these actions as governed, with built-in guardrails.
upsert_blueprint
Approval: see the ruleCreates a blueprint, or updates it when it already exists.
upsert_entity
Approval: see the ruleCreates an entity, or updates an existing one, merging by default so the fields you do not touch are kept.
upsert_scorecard
Approval: see the ruleCreates a scorecard, or updates it when it already exists. Port counts setting up scorecards among the jobs of its Builders, the platform engineers and admins who configure the portal.
upsert_action
Approval: see the ruleCreates a self-service action or automation, or updates it when it already exists.
What Claude asks before it acts
By default, Claude stops and asks for your go-ahead before any action it takes on an account on your behalf. The request shows up in the conversation at the moment it matters.
No source says whether Claude asks before run_action or an upsert_* tool, so that default rule covers all five. On Team and Enterprise workspaces, owners decide whether members may let certain actions through without being asked, and they can restrict what a connector may do for the whole organization, keeping reading and blocking changes, for example. On the Port side, your role and permissions decide which tools you get at all. Claude never sees more than you do.
Which plans include it
None of the 819 sheets in the official directory shows availability by plan. The answer, connector by connector, is not published anywhere: a real gap in the catalogue.
The general rule is published: remote connectors are open to all users on Claude, Cowork, Claude Desktop and mobile. On Team and Enterprise, an Owner or Primary Owner enables a connector for the organization before members can sign in. For the current state, check the Port IO sheet in the official directory.
Where this connector stops
A connector is not an automation. Claude calls these tools while it answers you: nothing starts when an entity changes, a scorecard drops or an integration fails.
This connector has no Claude help article of its own; what we know comes from the directory sheet and from Port's tool reference. The tool list is an observed floor, not a promise: an administrator can open actions that no public sheet shows. The directory's verified badge is not a security audit, and Anthropic writes on every sheet that it does not control which tools a publisher exposes. Port is the only source documenting this connector's tools.
Need help connecting Port IO to Claude?
A person reads every message.