Export chats from the Signal Desktop app to Markdown and HTML files with attachments. Each chat is exported as an individual .md/.html file and the attachments for each are stored in a separate folder. Attachments are linked from the Markdown files and displayed in the HTML (pictures, videos, voice notes).
Currently this seems to be the only way to get chat history out of Signal!
Adapted from mattsta/signal-backup, which I suspect will be hard to get working now.
An export for a group conversation looks as follows:
[2019-05-29, 15:04] Me: How is everyone?
[2019-05-29, 15:10] Aya: We're great!
[2019-05-29, 15:20] Jim: I'm not.
Images are attached inline with ![name](path)
while other attachments (voice notes, videos, documents) are included as links like [name](path)
so a click will take you to the file.
This is converted to HTML at the end so it can be opened with any web browser. The stylesheet .css
is still very basic but I'll get to it sooner or later.
This tool has some pretty difficult dependencies, so it's easier to get some help from Docker. For most people this will probably be the easiest way. It requires installing Docker and then pulling a 200MB image, so avoid this if data use is a concern.
First off, install Docker. And make sure you have Python installed.
Then install this package:
pip install signal-export
Then run the script! It will do some Docker stuff under the hood to get your data out of the encrypted database.
sigexport ~/signal-chats
# output will be saved to the supplied directory
NB On Windows/PowerShell, you may get an error like term 'sigexport' is not recognized
, in which case you can use the following:
python -m sigexport.main ~/signal-chats
Please fully exit your Signal app before proceeding, otherwise you will likely encounter an I/O disk
error, due to the message database being made read-only, as it was being accessed by the app.
See the full help info:
sigexport --help
Disable pagination on HTML, and overwrite anything at the destination:
sigexport --paginate=0 --overwrite ~/signal-chats
List available chats and exit:
sigexport --list-chats
Export only the selected chats:
sigexport --chats=Jim,Aya ~/signal-chats
You can add --source /path/to/source/dir/
if the script doesn't manage to find the Signal config location.
Default locations per OS are below.
The directory should contain a folder called sql
with db.sqlite
inside it.
- Linux:
~/.config/Signal/
- macOS:
~/Library/Application Support/Signal/
- Windows:
~/AppData/Roaming/Signal/
You can also use --old /previously/exported/dir/
to merge the new export with a previous one.
Nothing will be overwritten!
It will put the combined results in whatever output directory you specified and leave your previos export untouched.
Exercise is left to the reader to verify that all went well before deleting the previous one.
I don't recommend this, and you will have issues with file-ownership and other stuff. You can also run the Docker image directly, it just requires copy-pasting a much-longer command and being careful with volume mounts.
First set the appropriate environment variables for your OS:
# Only enter one of these!
SIGNAL_INPUT="$HOME/.config/Signal" # Linux
SIGNAL_INPUT="$HOME/Library/Application Support/Signal" # macOS
SIGNAL_INPUT="$HOME/AppData/Roaming/Signal" # Powershell
# And your output location (must be an absolute path)
SIGNAL_OUTPUT="$HOME/Downloads/signal-output"
Then run the below command, which pulls in the environment variables you set above.
# Note that the --overwrite flag is necessary when running like this
# careful where you point it!
docker run --rm \
-v "$SIGNAL_INPUT:/Signal" \
-v "$SIGNAL_OUTPUT:/output" \
carderne/sigexport:latest \
--overwrite /output \ # this line is obligatory!
--chats Jim # this line isn't
Running this script using the methods above requires you to trust that I haven't snuck anything into the Docker image.
You can inspect the code in this repo, and after pip install
ing, you can confirm that the code installed on your computer matches this repo.
But the methods above rely on the Docker image, which is a bit more complex.
You can check the Dockerfile in this repo, and the GitHub Actions workflow. You can check that the last Actions run matches the time on the last push to the DockerHub registry. But you can't guarantee that I didn't sneak something in, so the next thing is to check the image itself.
So you can run docker inspect carderne/sigexport
and check the Entrypoint
and Cmd
values.
Then you can run docker run --rm -it --entrypoint='' carderne/sigexport bash
and check what Entrypoint/Cmd values correspond to inside the container, and check that that matches the repo!
You can also just build your own Docker image from this repo:
git clone https://github.com/carderne/signal-export.git
cd signal-export
docker build -t yourname/sigexport .
And then run the Python script, but tell it to use the image you just created:
sigexport --docker-image yourname/sigexport outputdir/
This is hard mode, and involves installing more stuff. Probably easy on macOS, slightly involved on Linux, and impossible on Windows.
Before you can install signal-export
, you need to get sqlcipher
working.
Follow the instructions for your OS:
Install the required libraries.
sudo apt install libsqlite3-dev tclsh libssl-dev
Then clone sqlcipher and install it:
git clone https://github.com/sqlcipher/sqlcipher.git
cd sqlcipher
./configure --enable-tempstore=yes CFLAGS="-DSQLITE_HAS_CODEC" LDFLAGS="-lcrypto -lsqlite3"
make && sudo make install
- Install Homebrew.
- Run
brew install openssl sqlcipher
Ubuntu on WSL2 should work! That is, install WSL2 and Ubuntu on Windows, and then follow the For Linux instructions and feel your way forward. But probably just give up here and use the Docker method instead.
Then you're ready to install signal-export:
(Note the [sql]
that has been added!)
pip install signal-export[sql]
Then you should be able to use the Usage instructions as above.
git clone https://github.com/carderne/signal-export.git
cd signal-export
pip install -e .[dev,sql]
pre-commit install
Run tests with:
make test
And check types with:
mypy sigexport/
- signal-backup-decode might be easier if you use Android!
- signal2html also Android only