MCP App Store
by KeenEthics
Join icon

Join MCP Server for Claude and ChatGPT

By Join Inc.

Overview

Join aligns owners, design teams, and contractors within a unified system of record for preconstruction and the design phase. By integrating real-time cost and schedule data with design evolution, Join provides the decision intelligence necessary to manage risk on complex projects and ensure predictable project delivery at scale. This connector allows you to interact with your project- and company-level Join data from Claude.

How to connect

  1. Claude: Open the Join listing in the Claude connectors directory and choose Connect. Claude listing Join exposes 15 tools here.
  2. ChatGPT: Open the Join listing in the ChatGPT app directory and choose Connect. ChatGPT listing Join exposes 32 tools here.

Tools

add-comment-to-item

ChatGPT
Post a comment on a Join item. The comment is attributed to the authenticated user and is visible to everyone with access to the item.

assets-for-item

ChatGPT
List the files attached to a Join item or option, covering both those uploaded to Join and those linked from an external document system such as Procore, Autodesk, or Egnyte. File contents are not returned. Results are paginated. Item IDs are returned by items-for-project, and get-item reports how many assets an item has.

categorizations-for-project

ChatGPT
List the categorizations enabled on this Join project. Categorizations are the labeling schemes, such as trade, system, or building area, used to group and break down a project's estimate lines, items, or risks. PAGE output returns a page of categorizations with example categories. RESOURCE output stores every category of the project's categorizations, except the built-in MasterFormat and UniFormat, as a JSON table to download.

collaborators-for-project

ChatGPT
List the people collaborating on a Join project. Results are paginated.

comments-for-item

ChatGPT
List the discussion recorded on a Join item or option, oldest first. Results are paginated. Item IDs are returned by items-for-project, and get-item reports how many comments an item has.

companies-for-project

ChatGPT
List the companies working on a Join project. Results are paginated.

create-item

ChatGPT
Create a new item in a milestone of a Join project, by default the project's active milestone. The item is created as a private draft, visible only to the authenticated user until it is published from within Join.

create-project

ChatGPT
Create a new construction project in Join, owned by the authenticated user's company, along with its first milestone, which becomes the project's active milestone. Call this when a project that should be tracked in Join does not exist yet; check list-my-projects or search-projects first when it may already exist, since Join does not deduplicate by name and every call creates a separate project. Only the first milestone can be created here, so this is not the way to add a milestone to a project that already exists. The user's company must be entitled to create projects, otherwise the call fails. The status, design phase and delivery method are given by name; the accepted names are listed in this tool's input schema, and an unlisted name fails the call before anything is created. The project's type is given as an ID from list-project-types, which must be called first. The location is stored as free text with no coordinates. The result includes a url linking to the new project.

create-risk

ChatGPT
Record a new project-team-visible risk on a Join project, with an impact and a likelihood each rated from 1 to 5. The risk is visible to everyone with access to the project.

current-user

ChatGPT
Return the name of the Join user this session is authenticated as.

get-asset

ChatGPT
Get a file stored in Join by asset ID, such as an image listed by images-for-project. A PNG, JPEG, GIF, or WebP file comes back as an image; any other file comes back as an embedded resource. When the original is over 3 MiB, or the caller may see a file but not download its original, as with another company's logo, the file returned is its stored thumbnail, 200 pixels on the short edge, and thumbnail is true.

get-contingency-report

ChatGPT
Report the contingencies and allowances of a Join project as its Contingency & Allowance Report shows them: each one's starting amount in the milestone estimate, the adjustments incorporated from earlier milestones, the pending and accepted adjustments from items, and the amount remaining. Reports one milestone, with the items drawing on each contingency, or every milestone side by side. Requires the user's role on the project to view markups or owner costs.

get-detailed-milestone-report

ChatGPT
Get a detailed milestone report with categorization columns and estimate lines. This returns a resource URI that can be used to access the full report data. Requires the user's role on the project to view the cost details of every line asked for. A call the role cannot fully view is refused, and narrowing filters.sources, costMode.markupMode or costMode.includeOwnerCosts may make it viewable.

get-item

ChatGPT
Get the full detail of a single Join item or option, beyond what the list tools return. Item IDs are returned by items-for-project and search-items.

get-item-history

ChatGPT
List the edits recorded on Join items: what changed, who changed it, and when. An itemID reports that one item's edits; a projectID reports the edits across every item in that project. The window covers the entire history unless narrowed with since and until, and every kind of edit unless narrowed with eventTypes. Results are paginated, newest first. Item IDs are returned by items-for-project and search-items.

get-risk

ChatGPT
Get a single risk recorded on a Join project by its ID. Risk IDs are returned by risks-for-project.

images-for-project

ChatGPT
List the images a report about a Join project can use: the project's thumbnail, its print header, the logos of its owner, design team, and general contractor, and the logos of the companies on the project. Each image is identified by asset ID, and get-asset returns the file. printHeader lists what Join's own printed reports show in their header: the print header when the project has one, otherwise the team logos. A project that shows one of Join's default icons has no thumbnail entry.

items-for-project

ChatGPT
List the items and options in a Join project as of its active milestone by default. Results can be filtered, sorted, scoped to another milestone or to all milestones, and can include drafts. Items are the individual scope decisions tracked on a project. A visibility of PUBLISHED means the item is not a draft. Results are paginated. The item IDs it returns are the input to get-item, which carries an item's full detail.

list-benchmarking-reports

ChatGPT
List saved benchmarking reports, which compare cost and scope metrics across several Join projects, optionally filtered to those attached to one project. These, together with the reports saved on individual projects, are what users are referring to by "my reports" or "reports in Join". Each report includes a url linking to it in the Join web app. Reports saved against a single project are listed by list-user-reports.

list-my-projects

ChatGPT
List the construction projects in Join that the authenticated user is a member of (my) or that belong to their company (all). Each project includes a url linking to it in the Join web app. Results are paginated. The project IDs it returns are the input to the other project tools.

list-project-types

ChatGPT
List the project types a new project can be filed under, for the company the authenticated user belongs to. Join scopes project types to a company and groups them into categories, so two types can share a name under different categories and only the ID identifies one; that ID is what create-project's typeID argument takes.

list-user-reports

ChatGPT
List the reports saved against a single Join project, including milestone cost reports, variance reports, and other custom reports. These, together with benchmarking reports, are what users are referring to by "my reports" or "reports in Join". Reports are scoped to one project, and project IDs are returned by list-my-projects. Each report includes a url linking to it in the Join web app. Benchmarking reports that compare several projects are listed in full by list-benchmarking-reports.

metrics-for-project

ChatGPT
Get a Join project's metrics per milestone. Metrics are the project's enabled units of measure, such as GSF or beds, and each milestone records a quantity for some of them. A metric with no row at a milestone has no quantity entered there, which is not the same as zero.

milestones-for-project

ChatGPT
List the milestones of a Join project, optionally including drafts. Milestones correspond to new design and estimate issuances on the project. Results are paginated. Items and most cost reports are scoped to these milestoneIDs, and the activeMilestoneID is treated as the default.

project-costs

ChatGPT
Retrieve the total estimated cost of a Join project's active milestone, along with a breakdown of that total by cost category.

risks-for-project

ChatGPT
List the risks of a given type recorded on a Join project. Results are paginated.

search-items

ChatGPT
Keyword search over the items in the Join projects in scope, ranked by relevance. The query matches an item's number, name, and description, and the names of its categories, milestone, and project. PAGE output returns a page of matches in relevance order. RESOURCE output stores every match, up to 9,500, as JSON to download. Item IDs returned here are accepted by get-item.

search-projects

ChatGPT
Keyword search over the Join projects in scope, most recently updated first unless sortKey says otherwise. PAGE output returns a page of matches. RESOURCE output stores every match, up to 9,500, as JSON to download. The project IDs it returns are the input to the other project tools.

terminology-for-project

ChatGPT
Get the labels a Join project uses for its cost concepts, keyed by concept. A project can rename these concepts, and Join shows the renamed labels throughout that project. This server's other tools use the default labels. The concepts, with their default labels: ESTIMATE (Estimate) is the milestone's baseline estimate; items are changes relative to it. TARGET (Budget) is the target budget for the project. DELTA (Delta) is the budget minus the estimate: negative when the estimate is over budget, positive when under. RUNNING_TOTAL (Running Total) is the estimate plus accepted items, the project's current cost. GAP (Gap) is the budget minus the running total, the change still needed to meet the budget. DIRECT_COST (Direct Costs) is the cost of materials, labor, equipment, and similar, before markups. MARKUP (Markups) is indirect, below-the-line costs. COST_OF_CONSTRUCTION (Cost of Construction) is all project costs except owner costs. PROJECT_TOTAL (Project Total) is all project costs, including owner costs. PROJECT_RUNNING_TOTAL (Project Running Total) is the running total including owner costs.

timeline-for-project

ChatGPT
Return the scheduled activities that make up a Join project's timeline. The dependencies between activities are returned alongside them. Results are paginated.

update-item

ChatGPT
Update an existing Join item or option by ID. Omitted or empty arguments leave their field unchanged; a value is removed only through its clear flag. Supply at least one field or clear flag. Calls are idempotent, and the result names what actually changed alongside the item's stored values. A failed call can leave some fields already changed, and the error names them. The API rejects some status changes: an item with options moves only between ACCEPTED and INCORPORATED or among PENDING, REJECTED and NOTAPPLICABLE, and NOTCHOSEN applies only to an option whose parent is ACCEPTED or INCORPORATED.

update-project

ChatGPT
Update an existing Join project by ID. The call fails unless the authenticated user may edit the project's details, which is the same permission the web app gates these fields on. Omitted or empty arguments leave their field unchanged; the code is removed only through its clear flag. Supply at least one field or clear flag. Calls are idempotent, and the result names what actually changed alongside the project's stored values. The type is written by an API call of its own, so a failed call can leave some fields already changed, and the error names them; calling again from the same arguments finishes the job, since what is already stored is not written twice. The status is given by name, and the type as an ID from list-project-types.

add-comment-to-item

Claude

categorizations-for-project

Claude

create-item

Claude

create-risk

Claude

current-user

Claude

get-item

Claude

get-risk

Claude

items-for-project

Claude

list-benchmarking-reports

Claude

list-my-projects

Claude

list-user-reports

Claude

milestones-for-project

Claude

project-costs

Claude

risks-for-project

Claude

timeline-for-project

Claude

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/join"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/join)