Ithy Pro
Your research team. In your workflow.
Four independent AI answers and one combined response, returned as plain text in a single JSON response.
Simple by design
1,000 requests a month included with paid Pro. No streaming, SDK or extra subscription. Requests use the current Pro research team and sources; model updates follow the same registry as the website.
API answers stay private in your Ithy library. Your API allowance is separate from your web research answers. One request can run at a time per account.
Make your first API request
Send your key in the Authorization header and a unique Idempotency-Key for each question. Keep the same identifier if a connection times out: replaying it never starts another paid workflow. Allow up to 150 seconds. A running request returns 409; a completed replay returns the saved result.
curl https://ithy.com/api/v1/research \
-H "Authorization: Bearer $ITHY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: my-research-000001" \
-d '{"question":"How do heat pumps work in cold weather?"}'
The JSON response includes individual_answers, each with model and answer, plus combined_answer, sources and remaining quota. Answers are plain text, without HTML or Markdown formatting.
Questions: up to 2,000 characters. No attachments, images or videos. Your monthly request allowance is the API usage limit, with no additional dollar ceiling. Once paid research starts, that request uses a slot even if it fails or disconnects. Failures before dispatch return the slot. Replaying the same identifier never uses another slot or starts another paid job.
Building something useful? Tell us about it
