Pagination

Every list endpoint — list and list_results — is cursor-paginated. You almost never page by hand: the SDK's list methods return an iterator that fetches each page as you go.

Iterate automatically

# Walks every result across all pages, fetching more as needed.
for result in client.protein.design.list_results(design_id):
    print(result.id, result.metrics.binding_confidence)

# Async clients use `async for`.
async for result in client.protein.design.list_results(design_id):
    ...

Cursors and page fields

If you call the endpoints directly (or want explicit control), pagination is id-cursor based:

Field	Where	Meaning
limit	request	Max items per page (defaults to 100).
after_id	request	Return items after this id — pass the previous page's last_id to go forward.
before_id	request	Return items before this id — pass first_id to go backward.
data	response	The items in this page.
has_more	response	Whether another page exists.
last_id / first_id	response	Cursors for the next / previous page.

# First page, then the next one via the returned cursor.
curl "https://api.boltz.bio/compute/v1/protein/design/$RUN_ID/results?limit=100" \
  -H "x-api-key: $BOLTZ_API_KEY"

curl "https://api.boltz.bio/compute/v1/protein/design/$RUN_ID/results?limit=100&after_id=$LAST_ID" \
  -H "x-api-key: $BOLTZ_API_KEY"

Page manually in the SDK

When you need fine-grained control, the page object exposes the cursor directly:

page = client.protein.design.list_results(design_id)
print(page.data, page.last_id)

while page.has_next_page():
    page = page.get_next_page()
    print(len(page.data))

get_next_page() raises if there is no next page, so guard it with has_next_page() (use await on the async client).

admin.usage.list is the one exception: it uses an opaque page token (page request param, next_page in the response) instead of after_id / before_id. Auto-iteration works the same way.