POST · leads:write

Reply to a task

Answer a pending question or send a follow-up after the task finishes. Send answers or message.

Set up with your AI assistant

Paste the prompt into your coding assistant. It includes setup steps and a read-only connection check.

Request

Endpoint

POST https://leadsagent.io/api/v1/tasks/{id}/messages

Set LEADSAGENT_API_KEY in your local environment. Send it in the X-API-Key header.

Bash · API base URL

export LEADSAGENT_API_BASE="https://leadsagent.io/api/v1"
# For a local backend:
# export LEADSAGENT_API_BASE="http://127.0.0.1:8001/api/v1"

Bash · curl

curl --fail-with-body --silent --show-error --request POST "$LEADSAGENT_API_BASE/tasks/$TASK_ID/messages" -H "X-API-Key: $LEADSAGENT_API_KEY" -H "Content-Type: application/json" --data '{
  "answers": [
    "Yes, include independent practices"
  ]
}'

Replace TASK_ID with data.id from a task response. In PowerShell, use curl.exe and $env:LEADSAGENT_API_KEY. Save JSON in a file and pass --data-binary @request.json.

id (UUID): Required. The task ID.
answers (array of strings): Optional. Up to 20 answers in question order.
message (string): Optional. Follow-up instruction, up to 4,000 characters.

For a needs_input task, send one answer for each question.
For a finished task, send message to continue the same task.

Response

data is the updated task object. Follow-ups share the original task credit cap.

Example response

{
  "request_id": "6c4d0c26-8aa3-44b0-a50a-51b832b29465",
  "status": "success",
  "data": {
    "id": "73bd1ccb-1a44-4b41-84da-4ea7a5ad03f4",
    "kind": "agent",
    "status": "running",
    "prompt": "Find 20 dentists in Austin with emails",
    "source": "API key · research-example",
    "max_leads": 20,
    "created_at": "2026-10-21T12:00:00+00:00",
    "updated_at": "2026-10-21T12:00:00+00:00",
    "steps": [],
    "lead_count": 0,
    "results_ready": false,
    "status_url": "/api/v1/tasks/73bd1ccb-1a44-4b41-84da-4ea7a5ad03f4",
    "next": "Poll status_url for the new turn."
  },
  "meta": {}
}

Errors and retries explains status codes and retry rules. A successful request does not mean a background task has finished.

Read as Markdown