How to Connect Rippling to ChatGPT via MCP

💡 No install? Use cloud.anythingmcp.com directly. Sign in, click Connectors → Rippling, paste your credentials, mint an MCP API key and you're done. No Docker, no git clone, no local server.

Rippling + ChatGPT

Rippling HR platform through its REST API: workers and users, departments, teams, work locations, levels, titles, employment types and custom fields, compensation, leave types, balances and requests, supergroups and time entries. It also files and approves leave, records time, maintains the org structure and creates draft hires.

Prerequisites

See the full setup instructions baked into the connector (visible in the in-app store when you select the connector). The required environment variables for this connector are:

RIPPLING_API_TOKEN

Step 1 — Get credentials

Setting up takes about 5 minutes. You need a Rippling user with the permission Create and manage API Tokens and access to the data the AI should read.

1. In Rippling open Developer > API credentials (https://app.rippling.com/api-tokens/tokens) and click New API token.
2. Name the token and select its scopes. The token reaches only data that both its scopes and its owner's permission profile allow. Pick what you use:
   • companies.read (required, the install check), plus workers.read, users.read, departments.read, teams.read, work-locations.read, levels.read, titles.read, employment-types.read, job-codes.read, custom-fields.read, supergroups.read.
   • compensations.read only if the AI should see pay.
   • leave-types.read, leave-balances.read, leave-requests.read, and leave-requests.read-write to file or approve leave.
   • time-entries.read, and time-entries.read-write to record time (needs Rippling Time).
   • departments.read-write, teams.read-write, work-locations.read-write and draft-hires.read-write for the matching writes.
3. Copy the token right away (it is shown only once) into RIPPLING_API_TOKEN, without Bearer. Install the connector: rippling_list_companies is the install check.

Requests run as the token's owner. Rippling revokes a token that stays unused for 30 days and the tokens of an owner who is terminated, so pick an owner whose access is stable.

Personal data: workers, users and compensation packages hold personal and pay data: names, private emails, phone numbers, home addresses, date of birth, gender and salary. Ask only for the fields the task needs, do not copy them into other systems without a reason, and give the token only the scopes the team uses.

What it reads: the company (rippling_list_companies); workers, the employment records of employees and contractors, with status, work and personal email, start and end date, title, department, manager, level, teams, location, employment type and country (rippling_list_workers, rippling_get_worker); users, the login identity behind a worker's user_id (rippling_list_users, rippling_get_user); compensation packages with annual, monthly, weekly and hourly amounts, bonus targets and commission (rippling_list_compensations, and rippling_get_compensation with the worker's compensation_id); custom field definitions (rippling_list_custom_fields), whose values come with expand=custom_fields on workers; the org structure with departments, teams, work locations, levels, titles, employment types and job codes; supergroups and their members; leave types, leave balances in hours and days, and leave requests; time entries with job shifts, breaks, duration, status and pay period. expand (for example user,manager,department) embeds related objects in one call.

What it writes:

• Leave: rippling_create_leave_request files leave for a worker (PENDING follows the normal approval flow, APPROVED records leave that was already agreed); rippling_update_leave_request approves, rejects or cancels it through status, or changes its dates and comments.
• Time: rippling_create_time_entry records worked time as job shifts with start and end times or as a duration on a day; it takes an idempotency_key so a retry cannot create a duplicate. rippling_update_time_entry changes an entry; the API wants worker_id even when it does not change.
• Org structure: rippling_create_department, rippling_update_department, rippling_create_team, rippling_update_team, rippling_create_work_location and rippling_update_work_location. Updates change only the fields sent.
• Hiring: rippling_create_draft_hires creates up to five draft hires that go through Rippling's review and approval before anyone is hired.

Switched off at install: rippling_delete_team (workers lose the team assignment) and rippling_delete_time_entry. Neither can be undone; a workspace admin switches them on when the team needs them.

Filters: filter takes an OData-style expression on the fields each tool names, for example status eq 'ACTIVE', updated_at ge 2026-01-01T00:00:00 or status in ('HIRED','ACCEPTED'), combined with and and or. Strings take single quotes, and a filter has at most 64 nodes. Leave balances of one worker: worker_id eq '...'; open leave requests: status eq 'PENDING'.

Data formats: ids are 24-character hex strings. Dates are YYYY-MM-DD, timestamps ISO 8601 with offset. Money is { "currency_type": "USD", "value": 120000 } in major units. Worker status is INIT, HIRED, ACCEPTED, ACTIVE or TERMINATED. A field the token may not see comes back null and is named in __meta.redacted_fields with the reason, so null does not prove the field is empty: check the scopes and the owner's permissions first.

Paging: lists answer results and next_link, 50 items by default and up to 100 with limit. While next_link is not null, the AI calls again with cursor set to the cursor value inside next_link.

Limits and common errors: 600 requests per 10 seconds per endpoint and token, then 429 (wait a few seconds and retry). 400 means a bad filter, a field that cannot be expanded or limit over 100; 401 a wrong or revoked token; 403 a missing scope or a permission the owner lacks; 404 an unknown id.

The connector does not edit worker records or compensation, does not hire or terminate anyone directly and does not run payroll.

Step 2 — Install the adapter

curl -fsSL https://raw.githubusercontent.com/HelpCode-ai/anythingmcp/main/docker-compose.quickstart.yml -o docker-compose.yml
printf 'JWT_SECRET=%s\nENCRYPTION_KEY=%s\n' "$(openssl rand -hex 32)" "$(openssl rand -hex 32)" > .env
docker compose up -d

Step 3 — Add the connector in ChatGPT

Your server URL: in AnythingMCP, open MCP Servers → the server this connector is on, and copy its URL (https://cloud.anythingmcp.com/mcp/…). Use it wherever this guide shows YOUR_SERVER_ID.

ChatGPT adds MCP servers as plugins on its Plugins page, on the web and on paid plans. Developer mode is no longer needed. In a Business workspace only admins and owners can add one; in Enterprise and Edu an admin can grant the Create plugins with MCPs permission.

1. Go to chatgpt.com/plugins, click Add, then Add custom MCP server.
2. Enter a name, e.g. Name: Rippling.
3. Under Connection, enter the Server URL https://cloud.anythingmcp.com/mcp/YOUR_SERVER_ID and keep Authentication on OAuth.
4. Select I understand and want to continue, then Create as a plugin. When ChatGPT asks, sign in to AnythingMCP and allow access.
5. Install the new plugin from your personal plugins, then type @ in a chat and pick it.

Available tools

Example prompts

• "Who joined in the last 30 days? List name, title, department, manager and start date."
• "Show all pending leave requests with worker, dates, hours and leave type."
• "How much vacation does Maria Lopez have left, in days and in hours?"
• "Approve the pending leave request of Maria Lopez for 10 to 14 November."
• "How many active workers do we have per department and work location?"

FAQ

Does the free ChatGPT plan work with custom MCP servers? OpenAI documents custom MCP servers for paid plans only; the Free plan is not on its list.

Next steps

• Connector landing page
• Connect to Claude
• Connect to Cursor
• Connect to Gemini
• Connect to Meta Muse
• Connect to Microsoft Copilot