Transfer-Encoding vs Content-Encoding in HTTP

Transfer-Encoding vs Content-Encoding

These two HTTP headers are commonly confused but serve fundamentally different purposes.

Content-Encoding: About the Payload

Content-Encoding describes a transformation applied to the resource itself. The client stores the decoded version. It is an end-to-end header.

Content-Encoding: gzip

This means: "the body is the gzip-compressed form of the original resource."

Transfer-Encoding: About the Transport

Transfer-Encoding describes how the message body is formatted for transport between nodes. It is a hop-by-hop header that can be decoded by any intermediary.

Transfer-Encoding: chunked

This means: "the body is sent in chunks with size prefixes."

How They Interact

A response can use both simultaneously:

Content-Encoding: gzip
Transfer-Encoding: chunked

This means the original resource is gzip-compressed, and the compressed bytes are sent in chunks. This is common for dynamically generated, compressed responses where the server doesn’t know the total compressed size upfront.

Chunked Encoding Details

4\r\n        (chunk size in hex)
Wiki\r\n     (4 bytes of data)
6\r\n
pedia \r\n   (6 bytes of data)
0\r\n        (final chunk, size 0)
\r\n         (end of message)

Key Differences Summary

Aspect	Content-Encoding	Transfer-Encoding
Purpose	Resource transformation	Transport formatting
Scope	End-to-end	Hop-by-hop
Typical values	gzip, br, deflate	chunked
Cached?	Yes (encoded form)	No (decoded by proxy)
Affects Content-Length?	Yes	Replaces it