An open API and an MCP server for your whole workspace
A REST API with 370+ documented operations, webhooks for real-time events, and an MCP server that lets Claude, ChatGPT or Cursor work in your workspace with your permissions.
- 370+ API operations
- Webhooks with signing and resend
- MCP with OAuth 2.1
Does BusinessHRM have an API and an MCP server?
Yes. BusinessHRM has a REST API with 370+ documented operations across HR, attendance, payroll, projects, CRM, finance and support, with public docs and an OpenAPI file at app.businesshrm.com/developers. Each workspace also has an MCP server, so AI assistants such as Claude, ChatGPT and Cursor can read and update records with the connected user's own permissions.
What BusinessHRM does about it
BusinessHRM exposes the same records the app uses through a documented REST API, pushes changes out through webhooks as they happen, and offers a per-workspace MCP server for AI assistants. Every route checks the caller's role permissions, company scope and plan, so an API key or an AI connection can never see more than the person it belongs to.
Why teams go looking for API and MCP
- 01Data is locked inside a closed toolWithout an API, every report or sync means exporting a spreadsheet and importing it somewhere else by hand.
- 02Integrations poll and still miss changesScripts that check for updates every hour are slow and expensive, and they break when a field moves.
- 03AI tools cannot see your real dataPasting records into a chat window is slow, leaks more than it should and gives answers that are out of date the moment you close it.
Where AI can read, and where it can write
AI access follows the same permissions as the person using it, and writing is a deliberate choice.
Read and write through MCP
- Look up employees, attendance, leave, projects, tasks, contacts, deals, invoices and tickets the user may see
- Create and update records, such as a task, a contact or a leave request
- Ask for confirmation before deletes and sends, when the admin turns it on
Never
- See records outside the connected user's role or company
- Change data from the in-app AI chat, which is read-only
- Stay connected after an admin revokes the assistant
How API and MCP works
Connect in four steps
The same steps work for a script, an integration platform or an AI assistant.
- 1
Settings
Create an API key, or use OAuth
Admins create API keys in Settings. MCP clients can sign in with OAuth 2.1 instead, so no key is copied around.
- 2
Public docs
Read the reference
Every operation is documented at /developers with examples, an OpenAPI file and an llms.txt for AI coding tools.
- 3
About 99 events
Subscribe to webhooks
Choose the events you care about, add your URL and verify each delivery with the signing secret.
- 4
Streamable HTTP
Add the MCP server to your AI tool
Point Claude, ChatGPT or Cursor at your workspace's /mcp address and sign in as yourself.
Docs, keys and connected assistants
The developer docs are public; the settings screens are from a demo workspace.
Public API reference
Search every operation, see request and response examples, and download the OpenAPI file.
AI and integrations settings
Allow or block AI assistants for the company, require confirmation for deletes and sends, and see and revoke each connected assistant.
- Audit export as CSV
- Per-key access
Webhook guide
Event lists, payload examples and how to verify the signature, plus a delivery log with one-click resend in the app.
Ready-made connections
For common tools you may not need the API at all: Slack, Discord, GoHighLevel, QuickBooks, Xero, Google, Microsoft 365 and Zoom are in Connected apps.
Try the API on a free workspace
Create a workspace, add an API key and call the docs examples, or connect your AI assistant to the MCP server in a few minutes.
What you get with API & MCP
Included on every plan, including the free Starter plan for unlimited users.
370+ documented operations
Employees, attendance, leave, payroll, projects, tasks, time logs, contacts, deals, invoices, tickets, chat and more, all in one REST API.
OpenAPI and llms.txt
Generate a client from the OpenAPI file, or hand the llms.txt to an AI coding tool so it writes correct calls.
Permissions on every call
The API enforces the caller's role permissions, company scope and plan, so a key cannot reach further than its owner.
Webhooks with a delivery log
About 99 events, a signing secret you can rotate, test sends, and a log where any delivery can be resent.
MCP for AI assistants
A per-workspace MCP server over Streamable HTTP, with OAuth 2.1 sign-in (PKCE and dynamic client registration) or an API key.
Audit and revoke
Admins see every connected assistant, revoke access in one click and export an audit CSV.
What changes once it is running
- Your data goes where you need itSync to a warehouse, a BI tool or your own app without manual exports.
- React the moment something changesWebhooks push events as they happen, so integrations stop polling.
- AI that works on real recordsAsk an assistant to create a task or look up a balance, with the same rules as the app.
Who it is for
- DevelopersTeams building internal tools, syncs and dashboards on top of HR, project and finance data.
- AI-first teamsPeople who already work in Claude, ChatGPT or Cursor and want their workspace there too.
- Integrators and agenciesAnyone connecting BusinessHRM to client systems or automation platforms.
What people ask through MCP
Once connected, an assistant works with the same records and permissions as the person who signed in.
- Claude, ChatGPT or Cursor via MCP
"List everyone on leave next week and their leave type."
- Claude, ChatGPT or Cursor via MCP
"Create a contact for Ali Khan at Example Traders and add a follow-up deal in the Sales pipeline."
- Claude, ChatGPT or Cursor via MCP
"Which invoices are more than 30 days overdue, and who are the contacts?"
- In-app AI chat
"How many open tickets are assigned to me?"
Example prompts. The in-app chat answers questions and does not change records. Creating and updating records with AI happens through the MCP server, with the permissions of the person who connected it.
API & MCP questions, answered
Where are the API docs?
The REST API reference is public at app.businesshrm.com/developers, with request and response examples, an OpenAPI file and an llms.txt for AI coding tools.
How does API authentication work?
Admins create API keys in Settings. AI assistants connecting to the MCP server can sign in with OAuth 2.1 instead, or use an API key.
Can an API key see everything in the company?
No. Every call is checked against the role permissions, company and plan of the user the key or connection belongs to.
Which AI tools work with the MCP server?
Any MCP client that supports Streamable HTTP, including Claude, ChatGPT and Cursor. Each workspace has its own MCP address ending in /mcp.
Can AI change my data?
Through MCP, yes, within the connected user's permissions, and admins can require confirmation before deletes and sends. The in-app AI chat is read-only.
Are webhooks included?
Yes, on every plan. There are about 99 events with a signing secret, test sends, a delivery log and resend, plus native Slack and Discord formats.
Build on top of one source of truth
Start free, open the developer docs and connect your scripts, webhooks and AI tools to the same workspace your team uses.
Free forever for unlimited users. No credit card required.