The TenantWard API
Read your clients, issues, alerts, computers and the changes TenantWard made, from your own dashboards, PSA or scripts. It reads only; changes go through the agent door, where every one is checked against your change policy.
Get a token
An owner or admin makes one in Company › AI agents. Choose which clients it can see. The same token works for the API and for AI apps that connect over MCP. It stops working when you revoke it, or when the person who made it is no longer an owner or admin.
Send it with every request:
curl -H "Authorization: Bearer tw_agent_..." https://www.tenantward.com/api/v1/clients
Resources
Every resource is GET https://www.tenantward.com/api/v1/<resource> and returns { "data": … }.
| Resource | What it returns | Parameters |
|---|---|---|
| /whoami | The token: its company, name, the clients it can see and the most it may change. | None |
| /clients | Each client the token can see, with its connections, their scores and when each was last checked, and counts of open issues and alerts. | None |
| /issues | Open issues, worst first. |
|
| /alerts | Alerts, newest first. |
|
| /computers | Computers with the TenantWard agent, each with its last report and problems. |
|
| /changes | Changes TenantWard made, newest first, with whether each was undone. |
|
curl -H "Authorization: Bearer tw_agent_..." "https://www.tenantward.com/api/v1/issues?client=<client id>&limit=20"
Errors and limits
Errors come back as { "error": "…" } in plain words: 401 for a missing, revoked or expired token, 403 for a client the token can’t see, 400 for a bad parameter, 429 above 120 requests a minute.
The API needs a paid plan (Protect or above), like the agent door. Changing things (running a fix, undoing one) is through the MCP agent door for now, so every change keeps its backup, approval and undo.