Skip to main content

FIT file parser and decoder

Project description


Latest Docs

A FIT file parsing and decoding library written in Python3 (>= 3.6).

Usage Example

Read a FIT file, frame by frame:

import fitdecode

with fitdecode.FitReader('') as fit:
    for frame in fit:
        # The yielded frame object is of one of the following types:
        # * fitdecode.FitHeader (FIT_FRAME_HEADER)
        # * fitdecode.FitDefinitionMessage (FIT_FRAME_DEFINITION)
        # * fitdecode.FitDataMessage (FIT_FRAME_DATA)
        # * fitdecode.FitCRC (FIT_FRAME_CRC)

        if frame.frame_type == fitdecode.FIT_FRAME_DATA:
            # Here, frame is a FitDataMessage object.
            # A FitDataMessage object contains decoded values that
            # are directly usable in your script logic.

Command line utilities

fitjson exports JSON:

$ fitjson --pretty -o out_file.json

To ease the introspection or your FIT files, fittxt exports to a dedicated TXT format:

$ fittxt -o out_file.txt

Both commands accept a --filter option (or -f) which can be specified multiples times:

$ # include only RECORD messages:
$ fitjson -f=record -o out_file.json

$ # exclude FILE_ID and EVENT messages:
$ fitjson -f=-file_id -f=-event -o out_file.json


fitdecode is available on PyPI:

$ pip install fitdecode

Or, to get the latest working version, you can clone fitdecode’s source code repository before installing it:

$ git clone
$ cd fitdecode
$ python test     # optional step to run unit tests
$ python install

Note that for convenience, the cmd directory located at the root of the source code tree can safely be added to your PATH, so that fitdecode commands can be called without the package to be installed.


fitdecode is a non offensive and incompatible rewrite of the fitparse library, with some improvements and additional features, as well as efforts made to optimize both speed and memory usage.

Main differences between fitdecode and fitparse:

  • fitdecode’s API is not compatible with fitparse’s

  • fitdecode requires Python version 3.6 or greater

  • fitdecode is faster

  • fitdecode allows concurrent reading of multiple files by being thread-safe, in the sense that fitdecode’s objects keep their state stored locally

  • fitdecode does not discard the FIT header and the CRC footer while iterating a file, which allow to get a complete 1:1 representation of the file that is being read

  • This also allows the client to easily deal with so-called chained FIT files, as per FIT SDK definition (i.e. concatenated FIT files)

  • CRC computation and matching are both optional. CRC can be either matched, or only computed, or just ignored for faster reading.

  • fitdecode offers optional access to records, headers and footers in their binary form, to allow FIT file cutting, stitching and filtering at binary level

Why a new library?

A new library has been created instead of just offering to patch fitparse because many changes and adds in fitdecode break fitparse’s backward compatibilty and because it allowed more freedom during the development of fitdecode.


Documentation is available at


This project is distributed under the terms of the MIT license. See the LICENSE.txt file for details.


fitdecode is largely based on the generic approach adopted by fitparse to define FIT types and to decode raw values. That includes the module and all the classes it refers to, as well as the script

Change Log

v0.10.0 (2021-09-12)

  • fitjson: added --pretty option

  • fitjson: added --nounk option to filter-out unknown messages

  • fitjson: --filter option also allows to filter-out messages

  • fittxt: --filter option also allows to filter-out messages

  • fittxt: added --nounk option to filter-out unknown messages

  • Fixed: FitReader does not close a file-like object owned by the user

  • Fixed: FitReader.file_id gets reset upon FIT footer (CRC frame)

  • Fixed: utils.get_mesg_num() return value

  • Fixed: utils.get_mesg_field_num() return value

  • Minor corrections, improvements and code cleanup

v0.9.0 (2021-09-10)

  • FitReader gets new properties fit_file_index and fit_files_count

  • New CrcCheck policy: WARN

  • BREAKING CHANGE: CrcCheck default policy from RAISE to WARN

  • FitHeaderError exception messages a bit more helpful

  • Minor corrections and code cleanup

v0.8.0 (2021-09-09)

  • FitReader gets the error_handling argument to be less strict on malformed files (issues #13, #16, #18)

  • FIT SDK profile upgraded to v21.60

  • Minor corrections, improvements and cleanup on code and documentation

v0.7.0 (2020-10-04)

  • Compatibility with Apple Watch improved (issue #10)

  • FIT SDK profile upgraded to v21.38

  • generate_profile utility now supports recent SDK file structure

  • Minor improvements and cleanup on code and documentation

v0.6.0 (2019-11-02)

  • Added FitReader.last_timestamp property

  • Fixed: FitReader was raising KeyError instead of FitParseError when a dev_type was not found

  • FitParseError message contains more details upon malformed file in some cases

  • FIT SDK profile upgraded to v21.16

  • README’s usage example slightly improved

v0.5.0 (2019-04-11)

  • Added fitdecode.DataProcessorBase class

  • check_crc - the parameter to fitdecode.FitReader’s constructor - can now be either “enabled”, “read-only” or “disabled” (issue #1)

  • Minor speed improvements

v0.4.0 (2019-04-10)

  • Added fitdecode.FitDataMessage.has_field

  • fitdecode.FitDataMessage.get_fields is now a generator

  • fitdecode.FitDataMessage.get_values is now a generator

  • fitdecode.DefaultDataProcessor now converts hr.event_timestamp values that were populated from hr.event_timestamp_12 components to datetime.datetime objects for convenience

  • fitjson and fittxt utilities: * Added support for input files with Unicode characters * Still write output file even if an error occurred while parsing FIT file

  • Fixed handling of some FIT fields that are both scaled and components. See

  • Improved support for malformed FIT files. See

  • generate_profile utility slightly improved

  • Added some unit tests

  • Minor improvements and corrections

v0.3.0 (2018-07-27)

  • Added fitdecode.utils.get_mesg_field

  • Added fitdecode.utils.get_mesg_field_num

  • Minor improvements and corrections

v0.2.0 (2018-07-16)

  • Added FieldData.name_or_num

  • Added FitDataMessage.get_fields

  • Added FitDataMessage.get_values

  • Improved FitDataMessage.get_field (idx arg)

  • Improved FitDataMessage.get_value (idx arg)

  • Completed documentation of FitDataMessage

  • Improved documentation of FieldData

  • FitReader’s internal state is reset as well after a FitCRC has been yielded (i.e. not only when a FIT header is about to be read), in order to avoid incorrect behavior due to malformed FIT stream

v0.1.0 (2018-07-14)

  • Added class property frame_type (read-only) to FitHeader, FitCRC, FitDefinitionMessage and FitDataMessage (records module) to ease and speed up type checking

  • Added FitDataMessage.get_value method

  • string values with no null byte are still decoded (in full length)

  • cmd directory added to the source code tree for convenience

v0.0.1 (2018-07-08)

  • First release

v0.0.0 (2018-05-31)

  • Birth!

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

fitdecode-0.10.0.tar.gz (89.0 kB view hashes)

Uploaded source

Built Distribution

fitdecode-0.10.0-py3-none-any.whl (90.9 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page