Start a Run

Name: Rolli app
Address: 1212 5th Street, Suite 1-400,, Santa Monica, CA, 90401, US

Submit a plain-English question and the agent will plan, search, and produce a markdown intelligence report.

POST /api/v1/runs

Parameters

question	Required — the plain-English question (e.g., "What is being said about Tesla battery recalls this week?")
available_credits	Optional — reasoning effort for the run. `1` = low, `2` = medium, `3` = high, `4` = max (default). Higher values let the agent perform more searches and produce deeper analysis at higher cost.
mode	Optional — investigation mode. Default: `trend_briefing`. Allowed: `trend_briefing`, `monitoring`, `competitive_intel`.
time_window	Optional — time window to search. Default: `7d`. Examples: `24h`, `7d`, `30d`.
platforms	Optional — array of platforms to search. Allowed values: `twitter`, `reddit`, `bluesky`, `youtube`, `threads`, `facebook`.
max_post	Optional — maximum posts per platform. Default: 100.
parent_run_id	Optional — UUID of a parent run (for follow-ups).
metadata	Optional — arbitrary JSON object stored on the run for your own tracking.

Request Example

curl -X POST "https://agent.rolli.ai/api/v1/runs" \
  -H "X-ROLLI-TOKEN: your_token" \
  -H "X-ROLLI-USER-ID: your_user_id" \
  -H "Content-Type: application/json" \
  -d '{
  "question": "What is being said about Tesla battery recalls this week?",
  "mode": "trend_briefing",
  "time_window": "7d",
  "platforms": [
    "twitter",
    "reddit"
  ],
  "max_post": 100
}'

fetch("https://agent.rolli.ai/api/v1/runs", {
  method: "POST",
  headers: {
    "X-ROLLI-TOKEN": "your_token",
    "X-ROLLI-USER-ID": "your_user_id",
    "Content-Type": "application/json"
  },
  body: JSON.stringify({"question":"What is being said about Tesla battery recalls this week?","mode":"trend_briefing","time_window":"7d","platforms":["twitter","reddit"],"max_post":100})
})
  .then(res => res.json())
  .then(console.log);

require "net/http"
require "uri"
require "json"

uri = URI("https://agent.rolli.ai/api/v1/runs")
req = Net::HTTP::Post.new(uri)
req["X-ROLLI-TOKEN"] = "your_token"
req["X-ROLLI-USER-ID"] = "your_user_id"
req["Content-Type"] = "application/json"
req.body = {"question":"What is being said about Tesla battery recalls this week?","mode":"trend_briefing","time_window":"7d","platforms":["twitter","reddit"],"max_post":100}.to_json

res = Net::HTTP.start(uri.hostname, uri.port, use_ssl: true) do |http|
  http.request(req)
end

puts res.body

import requests
import json

url = "https://agent.rolli.ai/api/v1/runs"
headers = {
    "X-ROLLI-TOKEN": "your_token",
    "X-ROLLI-USER-ID": "your_user_id",
    "Content-Type": "application/json"
}

payload = {
  "question": "What is being said about Tesla battery recalls this week?",
  "mode": "trend_briefing",
  "time_window": "7d",
  "platforms": [
    "twitter",
    "reddit"
  ],
  "max_post": 100
}
response = requests.post(url, json=payload, headers=headers)

print(response.json())

Response Example

HTTP 202 Accepted. The run is queued — poll GET /api/v1/runs/{id} for progress and results.

{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "pending"
}

Status transitions: pending → running → completed (or failed / waiting).