Agent Computers API

An Agent Computer is a persistent workspace with a long-lived VM. Unlike task-scoped workspaces (which are provisioned per task), an Agent Computer persists across multiple tasks — all tasks share the same VM, filesystem, and installed packages.

Create Agent Computer

POST /agent-computers

Creates a new Agent Computer and starts provisioning its VM in the background. The response returns immediately with status: "provisioning" — poll GET /agent-computers/:id until status becomes "ready".

Request body:

Field	Type	Required	Description
name	string	No	Display name (default: "Agent Computer", max 200 chars)
icon	string	No	Icon identifier (max 50 chars)

curl -X POST https://api.rebyte.ai/v1/agent-computers \
  -H "API_KEY: rbk_xxx" \
  -H "Content-Type: application/json" \
  -d '{"name": "My Agent"}'

Response (201):

{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "name": "My Agent",
  "status": "provisioning",
  "sandboxId": null,
  "icon": null,
  "url": "https://app.rebyte.ai/agent/computer/550e8400-e29b-41d4-a716-446655440000",
  "provisionError": null,
  "createdAt": "2026-01-28T12:00:00.000Z",
  "updatedAt": "2026-01-28T12:00:00.000Z"
}

Agent Computer status:

Status	Description
provisioning	VM is being created
ready	VM is live and accepting tasks
error	Provisioning failed (see provisionError)

---

List Agent Computers

GET /agent-computers?limit=50&offset=0

Returns all Agent Computers for your organization, sorted by creation time (newest first).

Param	Type	Default	Description
limit	number	50	Results per page (1—100)
offset	number	0	Pagination offset

curl "https://api.rebyte.ai/v1/agent-computers" \
  -H "API_KEY: rbk_xxx"

Response:

{
  "data": [
    {
      "id": "550e8400-...",
      "name": "My Agent",
      "status": "ready",
      "sandboxId": "abc123",
      "icon": null,
      "url": "https://app.rebyte.ai/agent/computer/550e8400-...",
      "provisionError": null,
      "createdAt": "2026-01-28T12:00:00.000Z",
      "updatedAt": "2026-01-28T12:01:00.000Z"
    }
  ],
  "total": 3,
  "limit": 50,
  "offset": 0
}

---

Get Agent Computer

GET /agent-computers/:id

Returns full details for an Agent Computer including task count and recent tasks.

curl https://api.rebyte.ai/v1/agent-computers/550e8400-... \
  -H "API_KEY: rbk_xxx"

Response:

{
  "id": "550e8400-...",
  "name": "My Agent",
  "status": "ready",
  "sandboxId": "abc123",
  "icon": null,
  "url": "https://app.rebyte.ai/agent/computer/550e8400-...",
  "provisionError": null,
  "createdAt": "2026-01-28T12:00:00.000Z",
  "updatedAt": "2026-01-28T12:01:00.000Z",
  "taskCount": 15,
  "recentTasks": [
    {
      "id": "660e8400-...",
      "title": "Build REST API",
      "executor": "claude",
      "model": "claude-sonnet-4.6",
      "status": "completed",
      "createdAt": "2026-01-28T12:05:00.000Z",
      "completedAt": "2026-01-28T12:10:00.000Z"
    }
  ]
}

---

Running Tasks on an Agent Computer

To run a task on an existing Agent Computer, pass its id as the workspaceId when creating a task:

curl -X POST https://api.rebyte.ai/v1/tasks \
  -H "API_KEY: rbk_xxx" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Build a REST API with Express",
    "workspaceId": "550e8400-e29b-41d4-a716-446655440000"
  }'

This skips VM provisioning and is significantly faster than creating a new workspace. See the Tasks API for full task endpoint documentation.