Under the hoodTier: how much it should matter to you
2Useful: my rating, 1 to 5
3Signal: worth watching, 1 to 5
API RequestsArea: what it touches
Internal ChangesKind: in v2.1.311,
Internal ChangesSection of the release
What
Claude Code has a new way of building the request body it sends to the API. The request body is the data uploaded with each request, which includes the conversation.
A new name, api_incremental_request_body, joins the existing api_gzip_request_body entries in a list of API-related feature names.
The gzip path, which compresses request bodies before sending, now also accepts bodies prepared as binary buffers. Before, it compressed only text bodies.
It keeps fingerprints ("digests") of the blocks in a body from one request to the next. It also reports how much of each body was reused and whether an incremental body was sent.
Why
In a long conversation, much of each request repeats the previous one. Reusing those parts across requests can shrink what has to be uploaded each time.
