Follow-up tasks - Autumn AI

Every task has a durable task_id. To refine, expand, or fix results, continue the same task: the agent keeps the plan, the output rows, and the task-local files.

from autumn_sdk import AutumnClient

client = AutumnClient()

# Start a task
task = client.tasks.create("Find 20 AI infrastructure startups hiring founding engineers")

# Send more instructions to the same task
client.tasks.continue_(task.id, "Add each company's founder LinkedIn URL, and find 10 more rows")

# Read the result
result = client.tasks.output(task.id)
print(result["rows"])

import { AutumnClient } from "autumn-sdk";

const client = new AutumnClient();

const task = await client.tasks.create("Find 20 AI infrastructure startups hiring founding engineers");

await client.tasks.continue(task.task_id, "Add each company's founder LinkedIn URL, and find 10 more rows");

const result = await client.tasks.output(task.task_id);
console.log(result.rows);

When to continue vs. start fresh

Continue when the user references previous work: “add a column”, “find more”, “narrow to US-only”, “fix the empty fields”. Start a new task only for unrelated work. When you continue, rewrite the user’s request into a clear instruction stating the current goal, not just their raw words.

Execute and stop

tasks.execute(task_id) pushes an existing task toward producing output from its current plan.
tasks.stop(task_id) stops a run and returns the task to plan.

Human in the loop

Let the agent ask a clarifying question first.

Live messages

Watch each turn as it happens.