Protobuf media types: Media Types for Protocol Buffers
Official media type labels for serialized Protocol Buffer data
Kucherawy, M., Ed., Kumari, W., and R. Sloan, "Media Types for Protocol Buffers", RFC 9996, DOI 10.17487/RFC9996, July 2026, <https://www.rfc-editor.org/info/rfc9996>.
@misc{rfc9996,
series = {Request for Comments},
number = 9996,
howpublished = {RFC 9996},
publisher = {RFC Editor},
doi = {10.17487/RFC9996},
url = {https://www.rfc-editor.org/info/rfc9996},
author = {M. Kucherawy and W. Kumari and R. Sloan},
title = {{Media Types for Protocol Buffers}},
pagetotal = 14,
year = 2026,
month = jul,
} RFC text
Contents
Internet Engineering Task Force (IETF) M. Kucherawy, Ed.
Request for Comments: 9996
Category: Informational W. Kumari
ISSN: 2070-1721 R. Sloan
Google
July 2026
Media Types for Protocol Buffers
Abstract
This document registers media types for Protocol Buffers, a common
extensible mechanism for serializing structured data.
Status of This Memo
This document is not an Internet Standards Track specification; it is
published for informational purposes.
This document is a product of the Internet Engineering Task Force
(IETF). It represents the consensus of the IETF community. It has
received public review and has been approved for publication by the
Internet Engineering Steering Group (IESG). Not all documents
approved by the IESG are candidates for any level of Internet
Standard; see Section 2 of RFC 7841.
Information about the current status of this document, any errata,
and how to provide feedback on it may be obtained at
https://www.rfc-editor.org/info/rfc9996.
Copyright Notice
Copyright (c) 2026 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(https://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Revised BSD License text as described in Section 4.e of the
Trust Legal Provisions and are provided without warranty as described
in the Revised BSD License.
Table of Contents
1. Introduction
2. Key Words
3. Payload Description
4. Encoding Considerations
5. Versions
6. Security Considerations
7. IANA Considerations
7.1. Media Type: application/protobuf
7.2. Media Type: application/protobuf+json
8. References
8.1. Normative References
8.2. Informative References
Acknowledgments
Authors' Addresses
1. Introduction
Protocol Buffers (also called "Protobuf") were introduced in 2008 as
a free, open-source, platform-independent mechanism for transport and
storage of structured data. The use of Protobuf has become
increasingly common, and Protobuf implementations exist in many
languages (C++, C#, Dart, Go, Java, Kotlin, Objective-C, Python,
JavaScript, Ruby, Swift, and perhaps others). See [Protobuf] for
more information.
Protobuf consists of an interface definition language (IDL), wire
encoding formats, and language-specific implementations (typically
involving a generated API) so that clients and servers can be easily
deployed using a common schema. Protobuf supports two wire formats
for interchange: the format in [Binary] (which is optimized for wire
efficiency) and the format in [ProtoJSON] (which maps the Protobuf
schema onto a JSON structure).
Serialized objects are occasionally transported within media that
make use of media types (see [RFC2045]) to identify payloads.
Accordingly, current and historical media types used for this purpose
would benefit from registration. IANA has registered two Protobuf
media types; see Section 7.
2. Key Words
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
"OPTIONAL" in this document are to be interpreted as described in
BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all
capitals, as shown here.
3. Payload Description
The media types defined in this document are used in the transport of
serialized objects only. The IDL and object definitions, if
transported, would be used with any appropriate text media type. In
the three examples below, only the third (depicted in JSON) would
ever be used with these media types.
1. An example of using the IDL to specify a "Person" object:
edition = "2023";
message Person {
string name = 1;
int32 id = 2;
string email = 3;
}
2. An example of python code that uses code generated from the IDL
definition above to create an instance of a "Person" object:
person = Person()
person.id = 1234
person.name = "John Doe"
person.email = "jdoe@example.com"
3. An example of the above instance expressed in JSON:
{
"name": "John Doe",
"id": 1234,
"email": "jdoe@example.com"
}
4. Encoding Considerations
Protobuf supports the formats in [Binary] and [ProtoJSON] for
interchange, both of which are platform-independent. For binary
forms that need to transit non-binary transports, a base64 encoding
(e.g., [RFC4648]) is recommended.
Both of the media types defined in this document include an optional
"encoding" parameter indicating which encoding format is to be used
with that particular payload. This is included for future
extensibility. Valid values for this parameter are "binary" and
"json", and other values MUST be treated as an error. See Section 7
for the defaults for each of the two registered media types. Using
"binary" for the JSON type or "json" for the binary type MUST be
treated as an error.
5. Versions
[Proto2] was the first public version of the Protobuf schema
language; [Proto3], [Edition2023], and [Edition2024] came later, with
[Edition2024] being current at the time of writing. Future editions
of the IDL are expected.
These versions refer to evolutions of the schema of the IDL, not the
wire format. Accordingly, a serialized object generated by any of
these is compatible with any other. The media type registrations in
Section 7 include support for versioning of the wire format, should
it ever change, but do not refer to the IDL, which can evolve
independently.
Note that there may be semantic changes implicit in the IDL version,
which can affect the interpretation of otherwise compatible bits on
the wire. For example, in Proto2, unknown values of an enumeration
were interpreted as invalid, whereas in Proto3, they are retained.
Clients MUST reject payloads with an unsupported version number.
6. Security Considerations
The payload for these media types contains no directly executable
code. While it is common for a Protobuf definition to be used as
input to a code generator, which then produces something executable,
that applies to the schema language, not serializations.
Protobuf provides no security, privacy, integrity, or compression
services. Clients or servers for which this is a concern should
avail themselves of solutions that provide such capabilities (e.g.,
[RFC8446]). Implementations should be careful when processing
Protobuf like any binary format; for example, a malformed request to
a Protobuf server could be crafted to allocate a very large amount of
memory, potentially impacting other operations on that server.
Protobuf supports embedded content in string or bytes fields: In both
cases, applications should ensure that the format of the content is
precisely as expected. Note that UTF-8 validation of string fields
is optional (see [ProtoFeatures]), and a manual well-formedness check
may be necessary. Further, handling Unicode text generally can be
quite complex (due to the problems discussed in [UniChars] and
[RFC8264], for example), so it is best to rely on well-supported
internationalization libraries whenever possible.
In order to safely use Protobuf serializations on the Web, it is
important to ensure that they are not interpreted as another document
type, such as JavaScript: We recommend base64-encoding binary
Protobuf responses whenever possible to prevent parsing as active
content. Servers should generally follow the advice of [RFC9205] to
prevent content sniffing for all binary formats.
Further, when using JSON serializations, it is important that it is
clear to browsers that the content is pure JSON so that they can
inhibit Cross-Site Script Inclusion or side-channel attacks using
techniques such as Cross-Origin Read Blocking [CORB]. Per [RFC6839],
pure JSON content is indicated by a +json subtype suffix (see also
[MIMESNIFF]); thus, when serializing Protobuf content to JSON, users
MUST use the application/protobuf+json media type. When using JSON,
charset can prevent certain encoding confusion attacks, so users
should specify it for all JSON encodings.
In the type described in [Any], there is technically a link that was
intended to be dereferenced to obtain schemas for a given type;
however, this is not supported by widely used Protobuf
implementations.
7. IANA Considerations
As per the process defined in [RFC6838], IANA has registered
application/protobuf and application/protobuf+json in the "Media
Types" registry. The following are deprecated aliases: application/
x-protobuf, application/x-protobuffer, and application/
x-protobuf+json.
7.1. Media Type: application/protobuf
Type name: application
Subtype name: protobuf
Required parameters: N/A
Optional parameters:
encoding
Indicates the type of Protobuf encoding and is "binary" by
default for application/protobuf, indicating the format in
[Binary]. At the time of writing, no other encoding can be
used for application/protobuf, so this parameter is for
extensibility.
version
Indicates the version of the wire encoding specification (not
the schema language), with a default of 1. At the time of
writing, no Protobuf wire encodings are versioned, so this
parameter is for extensibility. Unversioned wire encodings
should be treated as having version 1.
Encoding considerations: binary
Security considerations: See Section 6 of RFC 9996.
Interoperability considerations: The Protobuf specification
[Protobuf] includes versioning provisions to ensure backward
compatibility when encountering payloads with unknown properties.
Published specification: [Protobuf]
Applications that use this media type: Any application with a need
to exchange or store structured objects across platforms or
implementations.
Fragment identifier considerations: N/A
Additional information:
Deprecated alias names for this type: application/x-protobuf,
application/x-protobuffer
Magic number(s): N/A
File extension(s): N/A
Macintosh file type code(s): N/A
Person & email address to contact for further information: Protobuf
Team <protobuf-team@google.com>
Intended usage: COMMON
Restrictions on usage: N/A
Author: Rob Sloan <rmsj@google.com>, Protobuf Team <protobuf-
team@google.com>
Change controller: IETF
7.2. Media Type: application/protobuf+json
Type name: application
Subtype name: protobuf+json
Required parameters: charset, which must be set to utf-8 (case-
insensitive)
Optional parameters:
encoding
Indicates the type of Protobuf encoding and is json by default
for application/protobuf+json, indicating the format in
[ProtoJSON]. At the time of writing, no other encoding can be
used for application/protobuf+json, so this parameter is for
extensibility.
version
Indicates the version of the wire encoding specification (not
the schema language), with a default of 1. At the time of
writing, no protobuf wire encodings are versioned, so this
parameter is for extensibility. Unversioned wire encodings
should be treated as having version 1.
Encoding considerations: Same as encoding considerations of
application/json as specified in [RFC8259], Section 11.
Security considerations: See Section 6 of RFC 9996.
Interoperability considerations: The Protobuf specification
[Protobuf] includes versioning provisions to ensure backward
compatibility when encountering payloads with unknown properties.
Published specification: [Protobuf]
Applications that use this media type: Any application with a need
to exchange or store structured objects across platforms or
implementations.
Fragment identifier considerations: N/A
Additional information:
Deprecated alias names for this type: application/x-protobuf+json
Magic number(s): N/A
File extension(s): N/A
Macintosh file type code(s): N/A
Person & email address to contact for further information: Protobuf
Team <protobuf-team@google.com>
Intended usage: COMMON
Restrictions on usage: N/A
Author: Rob Sloan <rmsj@google.com>, Protobuf Team <protobuf-
team@google.com>
Change controller: IETF
8. References
8.1. Normative References
[Protobuf] Google, LLC, "Protocol Buffers", <https://protobuf.dev/>.
[RFC2045] Freed, N. and N. Borenstein, "Multipurpose Internet Mail
Extensions (MIME) Part One: Format of Internet Message
Bodies", RFC 2045, DOI 10.17487/RFC2045, November 1996,
<https://www.rfc-editor.org/info/rfc2045>.
[RFC2119] Bradner, S., "Key words for use in RFCs to Indicate
Requirement Levels", BCP 14, RFC 2119,
DOI 10.17487/RFC2119, March 1997,
<https://www.rfc-editor.org/info/rfc2119>.
[RFC4648] Josefsson, S., "The Base16, Base32, and Base64 Data
Encodings", RFC 4648, DOI 10.17487/RFC4648, October 2006,
<https://www.rfc-editor.org/info/rfc4648>.
[RFC6838] Freed, N., Klensin, J., and T. Hansen, "Media Type
Specifications and Registration Procedures", BCP 13,
RFC 6838, DOI 10.17487/RFC6838, January 2013,
<https://www.rfc-editor.org/info/rfc6838>.
[RFC6839] Hansen, T. and A. Melnikov, "Additional Media Type
Structured Syntax Suffixes", RFC 6839,
DOI 10.17487/RFC6839, January 2013,
<https://www.rfc-editor.org/info/rfc6839>.
[RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
May 2017, <https://www.rfc-editor.org/info/rfc8174>.
[RFC8259] Bray, T., Ed., "The JavaScript Object Notation (JSON) Data
Interchange Format", STD 90, RFC 8259,
DOI 10.17487/RFC8259, December 2017,
<https://www.rfc-editor.org/info/rfc8259>.
8.2. Informative References
[Any] "any.proto Schema Definition", commit 97921a5, 12 February
2026,
<https://github.com/protocolbuffers/protobuf/blob/main/src/
google/protobuf/any.proto>.
[Binary] Google, LLC, "Protobuf Binary Wire Encoding Spec",
<https://protobuf.dev/programming-guides/encoding>.
[CORB] Chromium, "Cross-Origin Read Blocking for Web Developers",
<https://www.chromium.org/Home/chromium-security/corb-for-
developers>.
[Edition2023]
Google, LLC, "Protocol Buffers Edition 2023 Language
Specification", <https://protobuf.dev/reference/protobuf/
edition-2023-spec>.
[Edition2024]
Google, LLC, "Protocol Buffers Edition 2024 Language
Specification", <https://protobuf.dev/reference/protobuf/
edition-2024-spec>.
[MIMESNIFF]
WHATWG, "MIME Sniffing - MIME Type Groups", WHATWG Living
Standard,
<https://mimesniff.spec.whatwg.org/#mime-type-groups>.
Commit snapshot:
<https://mimesniff.spec.whatwg.org/commit-
snapshots/67bde18bc4f75a5c5b5dfb1217161137ce385124/>.
[Proto2] Google, LLC, "Protocol Buffers Language Specification
(Proto2 Syntax)",
<https://protobuf.dev/reference/protobuf/proto2-spec>.
[Proto3] Google, LLC, "Protocol Buffers Language Specification
(Proto3)",
<https://protobuf.dev/reference/protobuf/proto3-spec>.
[ProtoFeatures]
Google, LLC, "Protobuf Feature Settings for Editions",
<https://protobuf.dev/editions/features/>.
[ProtoJSON]
Google, LLC, "Protobuf JSON Wire Encoding Spec",
<https://protobuf.dev/programming-guides/json>.
[RFC8264] Saint-Andre, P. and M. Blanchet, "PRECIS Framework:
Preparation, Enforcement, and Comparison of
Internationalized Strings in Application Protocols",
RFC 8264, DOI 10.17487/RFC8264, October 2017,
<https://www.rfc-editor.org/info/rfc8264>.
[RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol
Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018,
<https://www.rfc-editor.org/info/rfc8446>.
[RFC9205] Nottingham, M., "Building Protocols with HTTP", BCP 56,
RFC 9205, DOI 10.17487/RFC9205, June 2022,
<https://www.rfc-editor.org/info/rfc9205>.
[UniChars] Bray, T. and P. Hoffman, "Unicode Character Repertoire
Subsets", RFC 9839, DOI 10.17487/RFC9839, August 2025,
<https://www.rfc-editor.org/info/rfc9839>.
Acknowledgments
Orie Steele provided valuable feedback to this work.
Authors' Addresses
Murray S. Kucherawy (editor)
Email: superuser@gmail.com
Warren Kumari
Google
Email: warren@kumari.net
Rob Sloan
Google
Email: varomodt@gmail.com
AI ELI5
On this page
§1What Protocol Buffers Are
Protocol Buffers (Protobuf) is a method developed by Google for serializing structured data—turning a complex object (like a person with a name, ID, and email) into a flat sequence of bytes that can be transmitted or stored, and then reconstructed into the original object. The Protobuf system includes three components: an interface definition language (IDL) for defining schemas, wire formats for encoding the data, and language-specific implementations in C++, Python, Java, Go, Swift, and many others.
Protobuf supports two wire formats: a binary format optimized for efficiency and a JSON format that represents the data as JSON. When serialized objects are transmitted over HTTP or stored in files, they need a media type label (also called a MIME type or content type) to tell receivers what format to expect. For decades, people used unofficial labels like application/x-protobuf; this RFC registers official types.
Protocol Buffers is a system for packing objects into bytes in a compact way. Like a shipping box needs a label saying what's inside, protocol buffer data needs a label saying 'this is Protocol Buffer format'.
Schema (IDL):
message Person {
string name = 1;
int32 id = 2;
string email = 3;
}
JSON encoding of an instance:
{
"name": "Jane Doe",
"id": 5678,
"email": "jane@example.com"
}§2The Two Wire Formats
The binary format encodes data as a compact sequence of bytes, optimized for network efficiency and storage size. Each field is encoded with a tag (a number identifying the field) and a wireformat (indicating the data type: int, string, bytes, etc.). This format is fast to serialize and deserialize but not human-readable.
The JSON format represents the same data as a JSON object, making it human-readable and easy to inspect with standard JSON tools. Both formats represent the same information; an application can choose either one depending on whether it prioritizes efficiency (binary) or readability/debuggability (JSON). The JSON format also makes it easier for web applications and JavaScript tools to work with the data.
Binary format is like a secret code—tiny and fast but nobody can read it without a decoder. JSON format is like English words—bigger but you can read it with any text editor.
§3Official Media Type Registrations
The RFC registers two media types with IANA. application/protobuf is used for binary Protocol Buffer data. It has optional parameters: encoding (defaults to 'binary', indicating the binary wire format) and version (defaults to 1, indicating the wire format version, for future extensibility). Clients that receive data with encoding=json or an unsupported version MUST treat it as an error.
application/protobuf+json is used for JSON-serialized data and REQUIRES a charset parameter set to utf-8 (case-insensitive). It also supports optional encoding (defaults to 'json') and version parameters. The +json suffix indicates that the content is pure JSON, which helps browsers and security tools properly handle the data and prevent Cross-Origin Read Blocking attacks.
The RFC also notes that application/x-protobuf, application/x-protobuffer, and application/x-protobuf+json are deprecated aliases that some implementations may still use, but new applications should use the official types.
Binary Protocol Buffer data is labeled 'application/protobuf'. JSON Protocol Buffer data is labeled 'application/protobuf+json'. Both can have optional settings for which version of the format is being used.
§4Schema Versions vs. Wire Format Versions
It is important to distinguish between the Protocol Buffer schema language version (Proto2, Proto3, Edition 2023, Edition 2024) and the wire format version. The schema language versions refer to the IDL syntax and features available when defining the schema. These evolve over time and bring semantic changes; for example, Proto3 handles unknown enumeration values differently than Proto2.
The wire format, however, has remained compatible across all schema versions. Any data serialized using any schema version will deserialize correctly with any other. Accordingly, the media type's version parameter refers only to the wire format version, not the schema language. The schema language can evolve independently. Future versions of the wire format may exist, but at the time of writing, all versions are wire format version 1.
Clients that receive a version parameter they do not understand MUST reject the payload.
The way you write the schema (how you define what fields a message has) has changed over time. But the way you encode it on the wire has stayed the same. If either one changes in the future, you label it with a version number.
§5Security and Safe Handling
Protocol Buffers themselves contain no executable code; they are data. However, several considerations apply. Embedded content: A Protobuf message can contain string or bytes fields with embedded data (like a JSON string inside a string field). Applications must ensure the embedded format is exactly what they expect and validate UTF-8 encoding if applicable. Memory allocation: Malformed Protobuf data can be crafted to request allocation of huge amounts of memory, potentially causing denial of service. Implementations should limit memory allocation.
Web security: When sending binary Protobuf data over the web, servers should base64-encode it to prevent browsers from interpreting it as JavaScript or other executable formats. When sending JSON-serialized Protobuf (application/protobuf+json), the +json suffix ensures browsers treat it as pure data, preventing Cross-Site Script Inclusion and Cross-Origin Read Blocking attacks. Charset should always be specified for JSON encodings.
Protobuf provides no encryption, authentication, or integrity checking. Applications requiring these should use TLS for transport and application-level encryption and signing as needed.
Protocol Buffer data is just bytes, not code. But a broken message could freeze a computer by asking for too much memory. On the web, send binary data as base64 to keep browsers from trying to run it as code.