List every run of a Sume Format in Python: the keyset cursor pager

To list all runs of one Format, call GET /v1/formats/{handle}/{slug}/runs with limit up to 100, and keep passing next_cursor as cursor until has_more is false. A Format that was never run returns an empty list, not an error.

The cursor is a keyset over created_at and id, so new runs arriving during a walk do not shift the pages the way an offset would.

The response shape

From the Runs and results page (read 2026-10-03).

A pager

Store the cursor only as an opaque string and never parse it. Each item in data is a run receipt, so you can read status and usage without a second call.

import json, os, urllib.parse, urllib.request

BASE = "https://api.sume.com/v1/formats/acme/teaser/runs"
KEY = os.environ["SUME_API_KEY"]

def pages():
    cursor = None
    while True:
        q = {"limit": 100}
        if cursor:
            q["cursor"] = cursor
        req = urllib.request.Request(
            BASE + "?" + urllib.parse.urlencode(q),
            headers={"Authorization": "Bearer " + KEY})
        page = json.load(urllib.request.urlopen(req))
        yield from page["data"]
        if not page["has_more"]:
            return
        cursor = page["next_cursor"]

print(sum(1 for _ in pages()))

Reads count against your budget

Listing is a GET, so it spends the read budget, which is forty times the write budget on every plan. A walk of a thousand runs is ten calls. For the arithmetic see how the other list endpoints page in the jobs list cursor, and for totalling cost from the receipts you just listed, cost per row from receipts.