Your app should open the ticket, not a person

PimiDesk is an email helpdesk with a REST API for everything the dashboard does. Your app creates tickets, your tools hear about changes through webhooks, and your team answers from one inbox.

The week without it

A payment fails and nobody tells support. An error tracker catches a crash, and someone copies it into the helpdesk by hand. A customer writes in, and the person answering has to ask which account they mean.

Your app creates the ticket

When your code sees a failed payment or a crash, it opens a ticket with the customer attached. One request, with an API token. The contact is found by email or created.

curl -X POST \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "subject": "Payment processing failed",
    "description": "Automated alert: payment #12345 failed with error code DECLINED",
    "contact": { "name": "Maya Ross", "email": "[email protected]" },
    "priority": "high"
  }' \
  https://desk-api.pimi.io/api/v1/workspaces/WORKSPACE_ID/tickets

Creating tickets through the API needs a paid plan: a trial workspace can only reply to mail it receives. The API reference lists all 250+ operations.

Your tools hear about it

Webhooks cover 17 event types, and every delivery is signed.

• A high-priority ticket is created: post it to your engineering channel
• A ticket is closed: update the issue it came from
• A response target is missed: start your escalation

A workspace per product

Each workspace has its own addresses, contacts, tags and members, so one product's customers never show up in another's queue. Inside a workspace, an address feeds tickets, a shared inbox, or one person's private inbox.

Response targets per priority

Set a first-response and a resolution target for low, medium and high priority. When an assigned ticket misses one, the assignee gets an email and the sla.breached webhook fires. Targets run on clock time, and there is no warning before a breach. The SLA guide has the rules.

For response and resolution times per member, ask your AI tool: the MCP server has a statistics tool. The dashboard has no report page for it.

Get started