POST/api/v1/ask

Ask Markdrop

Answers a question from everything you can open (meetings and their transcripts, documents, pages, files and Guides) with Claude Haiku 5.5, citing sources as [n] and meeting transcript moments as [n-sK]. Pass thread_id to ask a follow-up in a conversation, and intent catch_up (the last 7 days) or my_actions (your open tasks) for the ready-made questions; question may be empty with an intent. Every cited item is checked against your access again before it's returned. Streams server-sent events: status (the stage), results (the relevant items), delta (answer text as it's written), then done with answer, citations, ranked and thread_id; or limit when an AI limit stops it, or error. Counts towards your AI search limits.

API token required30 per minute per IPMCP: ask_documents

Guide: Ask Markdrop

Code samples and Try it

Request body

application/json, required

questionstring
Up to 500 characters. Default "".
thread_idstring
Up to 40 characters.
intent"ask" | "catch_up" | "my_actions"

Responses

200OK.
401No token, or the token is revoked or mistyped.
422The request didn't validate. detail lists each problem, with where it is (loc) and why (msg).
429Too many requests from your IP address. Wait the seconds in Retry-After.

const res = await fetch("https://api.markdrop.in/api/v1/ask", { method: "POST", headers: { Authorization: `Bearer ${process.env.MARKDROP_TOKEN}`, "Content-Type": "application/json", }, body: JSON.stringify({ "question": "When do we ship the release?" }), }); const data = await res.text();

import os import requests r = requests.post( "https://api.markdrop.in/api/v1/ask", headers={"Authorization": f"Bearer {os.environ['MARKDROP_TOKEN']}"}, json={"question": "When do we ship the release?"}, ) r.raise_for_status() print(r.text)