Skip to main content
Threads persist context across multiple instructions to the same computer. The AI remembers what it did previously so you can build on prior steps. Threads are created automatically when you call the Completions endpoint. The thread_id is returned in the response.

​
List Threads

GET /api/v1/threads?computer_id={computer_id}

{
  "object": "list",
  "data": [
    {
      "id": "thr_abc123",
      "object": "thread",
      "created_at": "2026-03-14T10:00:00Z",
      "title": "Firefox AI search",
      "message_count": 4
    }
  ]
}

​
Get Thread

GET /api/v1/threads/{thread_id}
Returns the thread with its full message history. 
{
  "id": "thr_abc123",
  "object": "thread",
  "created_at": "2026-03-14T10:00:00Z",
  "title": "Firefox AI search",
  "messages": [
    {"role": "user", "content": "Open Firefox and search for AI news"},
    {"role": "assistant", "content": "Done. Firefox shows the search results."}
  ]
}

​
Create Thread

POST /api/v1/threads

{"computer_id": "your-computer-id"}

​
Delete Thread

DELETE /api/v1/threads/{thread_id}

​
Example

from orgo import Computer

computer = Computer()

# Step 1
result = computer.prompt("Open Firefox and go to github.com")

# Step 2 — AI remembers it already opened Firefox
result2 = computer.prompt("Search for 'orgo'", thread_id=result["thread_id"])

# Step 3
result3 = computer.prompt("Click the first result", thread_id=result["thread_id"])