---
title: "Start Expert Desktop"
method: POST
path: "/api/experts/{expert_id}/computer/desktop"
tags: ["v2", "experts", "experts", "private"]
---

# Start Expert Desktop

`POST /api/experts/{expert_id}/computer/desktop`

Start or resume the expert's desktop and return its live stream.

This is the same box the expert's next ``start_desktop`` turn reconnects
to, so what the user does here is what the expert sees.

## Path parameters

- `expert_id` string, required

## Response `200`

Successful Response

- DesktopStream
  - `kind` 'desktop_stream'
  - `url` string, required
  - `provider` 'e2b'
  - `sandbox_id` string, required
  - `requires_auth` boolean

## Other responses

- `401` — Authentication required
- `404` — Expert not found
- `422` — Validation Error
- `502` — The desktop could not be started
- `503` — E2B is not configured

## Changes

- **2026-09-16** `31650764b5ef` — 1 info
  - endpoint added

[Change history](https://skmtc.dev/significant-gravitas/apis/autogpt-agent-server/changes/api/experts/:expert_id/computer/desktop/post.md)

---

[API](https://skmtc.dev/significant-gravitas/apis/autogpt-agent-server.md) · [All operations](https://skmtc.dev/significant-gravitas/apis/autogpt-agent-server/llms.txt) · [OpenAPI document](https://skmtc.dev/significant-gravitas/apis/autogpt-agent-server/revisions/5f27fa9d6c03?raw)