Paperless provides a wide range of customizations. Depending on how you run paperless, these settings have to be defined in different places.

  • If you run paperless on docker, paperless.conf is not used. Rather, configure paperless by copying necessary options to docker-compose.env.

  • If you are running paperless on anything else, paperless will search for the configuration file in these locations and use the first one it finds:


Required services


This is required for processing scheduled tasks such as email fetching, index optimization and for training the automatic document matcher.

Defaults to redis://localhost:6379.


By default, sqlite is used as the database backend. This can be changed here. Set PAPERLESS_DBHOST and PostgreSQL will be used instead of mysql.


Adjust port if necessary.

Default is 5432.


Database name in PostgreSQL.

Defaults to “paperless”.


Database user in PostgreSQL.

Defaults to “paperless”.


Database password for PostgreSQL.

Defaults to “paperless”.


SSL mode to use when connecting to PostgreSQL.

See the official documentation about sslmode.

Default is prefer.

Paths and folders


This where your documents should go to be consumed. Make sure that it exists and that the user running the paperless service can read/write its contents before you start Paperless.

Don’t change this when using docker, as it only changes the path within the container. Change the local consumption directory in the docker-compose.yml file instead.

Defaults to “../consume/”, relative to the “src” directory.


This is where paperless stores all its data (search index, SQLite database, classification model, etc).

Defaults to “../data/”, relative to the “src” directory.


Instead of removing deleted documents, they are moved to this directory.

This must be writeable by the user running paperless. When running inside docker, ensure that this path is within a permanent volume (such as “../media/trash”) so it won’t get lost on upgrades.

Defaults to empty (i.e. really delete documents).


This is where your documents and thumbnails are stored.

You can set this and PAPERLESS_DATA_DIR to the same folder to have paperless store all its data within the same volume.

Defaults to “../media/”, relative to the “src” directory.


Override the default STATIC_ROOT here. This is where all static files created using “collectstatic” manager command are stored.

Unless you’re doing something fancy, there is no need to override this.

Defaults to “../static/”, relative to the “src” directory.


Changes the filenames paperless uses to store documents in the media directory. See File name handling for details.

Default is none, which disables this feature.


This is where paperless will store log files.

Defaults to “PAPERLESS_DATA_DIR/log/”.



Maximum file size for log files before they are rotated, in bytes.

Defaults to 1 MiB.


Number of rotated log files to keep.

Defaults to 20.

Hosting & Security


Paperless uses this to make session tokens. If you expose paperless on the internet, you need to change this, since the default secret is well known.

Use any sequence of characters. The more, the better. You don’t need to remember this. Just face-roll your keyboard.

Default is listed in the file src/paperless/


This setting can be used to set the three options below (ALLOWED_HOSTS, CORS_ALLOWED_HOSTS and CSRF_TRUSTED_ORIGINS). If the other options are set the values will be combined with this one. Do not include a trailing slash. E.g.

Defaults to empty string, leaving the other settings unaffected.


A list of trusted origins for unsafe requests (e.g. POST). As of Django 4.0 this is required to access the Django admin via the web. See

Can also be set using PAPERLESS_URL (see above).

Defaults to empty string, which does not add any origins to the trusted list.


If you’re planning on putting Paperless on the open internet, then you really should set this value to the domain name you’re using. Failing to do so leaves you open to HTTP host header attacks:

Just remember that this is a comma-separated list, so “” is fine, as is “,”, but NOT ”” or “,”

Can also be set using PAPERLESS_URL (see above).

If manually set, please remember to include “localhost”. Otherwise docker healthcheck will fail.

Defaults to “*”, which is all hosts.


You need to add your servers to the list of allowed hosts that can do CORS calls. Set this to your public domain name.

Can also be set using PAPERLESS_URL (see above).

Defaults to “http://localhost:8000”.


To host paperless under a subpath url like you set this value to /paperless. No trailing slash!

Defaults to none, which hosts paperless at “/”.


Override the STATIC_URL here. Unless you’re hosting Paperless off a subdomain like /paperless/, you probably don’t need to change this.

Defaults to “/static/”.


Specify a username here so that paperless will automatically perform login with the selected user.


Do not use this when exposing paperless on the internet. There are no checks in place that would prevent you from doing this.

Defaults to none, which disables this feature.


If this environment variable is specified, Paperless automatically creates a superuser with the provided username at start. This is useful in cases where you can not run the createsuperuser command separately, such as Kubernetes or AWS ECS.



This will not change an existing [super]user’s password, nor will it recreate a user that already exists. You can leave this throughout the lifecycle of the containers.


(Optional) Specify superuser email address. Only used when PAPERLESS_ADMIN_USER is set.

Defaults to root@localhost.


Only used when PAPERLESS_ADMIN_USER is set. This will be the password of the automatically created superuser.


Specify a prefix that is added to the cookies used by paperless to identify the currently logged in user. This is useful for when you’re running two instances of paperless on the same host.

After changing this, you will have to login again.

Defaults to "", which does not alter the cookie names.


Allows authentication via HTTP_REMOTE_USER which is used by some SSO applications.


This will allow authentication by simply adding a Remote-User: <username> header to a request. Use with care! You especially must ensure that any such header is not passed from your proxy server to paperless.

If you’re exposing paperless to the internet directly, do not use this.

Also see the warning in the official documentation <>.

Defaults to false which disables this feature.


If PAPERLESS_ENABLE_HTTP_REMOTE_USER is enabled, this property allows to customize the name of the HTTP header from which the authenticated username is extracted. Values are in terms of [HttpRequest.META]( Thus, the configured value must start with HTTP_ followed by the normalized actual header name.



URL to redirect the user to after a logout. This can be used together with PAPERLESS_ENABLE_HTTP_REMOTE_USER to redirect the user back to the SSO application’s logout page.

Defaults to None, which disables this feature.

OCR settings

Paperless uses OCRmyPDF for performing OCR on documents and images. Paperless uses sensible defaults for most settings, but all of them can be configured to your needs.


Customize the language that paperless will attempt to use when parsing documents.

It should be a 3-letter language code consistent with ISO 639:

Set this to the language most of your documents are written in.

This can be a combination of multiple languages such as deu+eng, in which case tesseract will use whatever language matches best. Keep in mind that tesseract uses much more cpu time with multiple languages enabled.

Defaults to “eng”.

Note: If your language contains a ‘-’ such as chi-sim, you must use chi_sim


Tell paperless when and how to perform ocr on your documents. Four modes are available:

  • skip: Paperless skips all pages and will perform ocr only on pages where no text is present. This is the safest option.

  • skip_noarchive: In addition to skip, paperless won’t create an archived version of your documents when it finds any text in them. This is useful if you don’t want to have two almost-identical versions of your digital documents in the media folder. This is the fastest option.

  • redo: Paperless will OCR all pages of your documents and attempt to replace any existing text layers with new text. This will be useful for documents from scanners that already performed OCR with insufficient results. It will also perform OCR on purely digital documents.

    This option may fail on some documents that have features that cannot be removed, such as forms. In this case, the text from the document is used instead.

  • force: Paperless rasterizes your documents, converting any text into images and puts the OCRed text on top. This works for all documents, however, the resulting document may be significantly larger and text won’t appear as sharp when zoomed in.

The default is skip, which only performs OCR when necessary and always creates archived documents.

Read more about this in the OCRmyPDF documentation.


Tells paperless to use unpaper to clean any input document before sending it to tesseract. This uses more resources, but generally results in better OCR results. The following modes are available:

  • clean: Apply unpaper.

  • clean-final: Apply unpaper, and use the cleaned images to build the output file instead of the original images.

  • none: Do not apply unpaper.

Defaults to clean.


clean-final is incompatible with ocr mode redo. When both clean-final and the ocr mode redo is configured, clean is used instead.


Tells paperless to correct skewing (slight rotation of input images mainly due to improper scanning)

Defaults to true, which enables this feature.


Deskewing is incompatible with ocr mode redo. Deskewing will get disabled automatically if redo is used as the ocr mode.


Tells paperless to correct page rotation (90°, 180° and 270° rotation).

If you notice that paperless is not rotating incorrectly rotated pages (or vice versa), try adjusting the threshold up or down (see below).

Defaults to true, which enables this feature.


Adjust the threshold for automatic page rotation by PAPERLESS_OCR_ROTATE_PAGES. This is an arbitrary value reported by tesseract. “15” is a very conservative value, whereas “2” is a very aggressive option and will often result in correctly rotated pages being rotated as well.

Defaults to “12”.


Specify the the type of PDF documents that paperless should produce.

  • pdf: Modify the PDF document as little as possible.

  • pdfa: Convert PDF documents into PDF/A-2b documents, which is a subset of the entire PDF specification and meant for storing documents long term.

  • pdfa-1, pdfa-2, pdfa-3 to specify the exact version of PDF/A you wish to use.

If not specified, pdfa is used. Remember that paperless also keeps the original input file as well as the archived version.


Tells paperless to use only the specified amount of pages for OCR. Documents with less than the specified amount of pages get OCR’ed completely.

Specifying 1 here will only use the first page.

When combined with PAPERLESS_OCR_MODE=redo or PAPERLESS_OCR_MODE=force, paperless will not modify any text it finds on excluded pages and copy it verbatim.

Defaults to 0, which disables this feature and always uses all pages.


Paperless will OCR any images you put into the system and convert them into PDF documents. This is useful if your scanner produces images. In order to do so, paperless needs to know the DPI of the image. Most images from scanners will have this information embedded and paperless will detect and use that information. In case this fails, it uses this value as a fallback.

Set this to the DPI your scanner produces images at.

Default is none, which will automatically calculate image DPI so that the produced PDF documents are A4 sized.


Paperless will not OCR images that have more pixels than this limit. This is intended to prevent decompression bombs from overloading paperless. Increasing this limit is desired if you face a DecompressionBombError despite the concerning file not being malicious; this could e.g. be caused by invalidly recognized metadata. If you have enough resources or if you are certain that your uploaded files are not malicious you can increase this value to your needs. The default value is 256000000, an image with more pixels than that would not be parsed.


OCRmyPDF offers many more options. Use this parameter to specify any additional arguments you wish to pass to OCRmyPDF. Since Paperless uses the API of OCRmyPDF, you have to specify these in a format that can be passed to the API. See the API reference of OCRmyPDF for valid parameters. All command line options are supported, but they use underscores instead of dashes.


Paperless has been tested to work with the OCR options provided above. There are many options that are incompatible with each other, so specifying invalid options may prevent paperless from consuming any documents.

Specify arguments as a JSON dictionary. Keep note of lower case booleans and double quoted parameter names and strings. Examples:

{"deskew": true, "optimize": 3, "unpaper_args": "--pre-rotate 90"}

Tika settings

Paperless can make use of Tika and Gotenberg for parsing and converting “Office” documents (such as “.doc”, “.xlsx” and “.odt”). If you wish to use this, you must provide a Tika server and a Gotenberg server, configure their endpoints, and enable the feature.


Enable (or disable) the Tika parser.

Defaults to false.


Set the endpoint URL were Paperless can reach your Tika server.

Defaults to “http://localhost:9998”.


Set the endpoint URL were Paperless can reach your Gotenberg server.

Defaults to “http://localhost:3000”.

If you run paperless on docker, you can add those services to the docker-compose file (see the provided docker-compose.sqlite-tika.yml file for reference). The changes requires are as follows:

    # ...

        # ...

            # ...

            PAPERLESS_TIKA_GOTENBERG_ENDPOINT: http://gotenberg:3000
            PAPERLESS_TIKA_ENDPOINT: http://tika:9998

    # ...

        image: gotenberg/gotenberg:7.4
        restart: unless-stopped
            - "gotenberg"
            - "--chromium-disable-routes=true"

        restart: unless-stopped

Add the configuration variables to the environment of the webserver (alternatively put the configuration in the docker-compose.env file) and add the additional services below the webserver service. Watch out for indentation.

Make sure to use the correct format PAPERLESS_TIKA_ENABLED = 1 so python_dotenv can parse the statement correctly.

Software tweaks


Paperless does multiple things in the background: Maintain the search index, maintain the automatic matching algorithm, check emails, consume documents, etc. This variable specifies how many things it will do in parallel.


Furthermore, paperless uses multiple threads when consuming documents to speed up OCR. This variable specifies how many pages paperless will process in parallel on a single document.


Ensure that the product


does not exceed your CPU core count or else paperless will be extremely slow. If you want paperless to process many documents in parallel, choose a high worker count. If you want paperless to process very large documents faster, use a higher thread per worker count.

The default is a balance between the two, according to your CPU core count, with a slight favor towards threads per worker:

CPU core count
























If you only specify PAPERLESS_TASK_WORKERS, paperless will adjust PAPERLESS_THREADS_PER_WORKER automatically.


Machines with few cores or weak ones might not be able to finish OCR on large documents within the default 1800 seconds. So extending this timeout may prove to be useful on weak hardware setups.


If PAPERLESS_WORKER_TIMEOUT has been configured, the retry time for a task can also be configured. By default, this value will be set to 10s more than the worker timeout. This value should never be set less than the worker timeout.


Set the time zone here. See for details on how to set it.

Defaults to UTC.


If paperless won’t find documents added to your consume folder, it might not be able to automatically detect filesystem changes. In that case, specify a polling interval in seconds here, which will then cause paperless to periodically check your consumption directory for changes. This will also disable listening for file system changes with inotify.

Defaults to 0, which disables polling and uses filesystem notifications.


When the consumer detects a duplicate document, it will not touch the original document. This default behavior can be changed here.

Defaults to false.


Enable recursive watching of the consumption directory. Paperless will then pickup files from files in subdirectories within your consumption directory as well.

Defaults to false.


Set the names of subdirectories as tags for consumed files. E.g. <CONSUMPTION_DIR>/foo/bar/file.pdf will add the tags “foo” and “bar” to the consumed file. Paperless will create any tags that don’t exist yet.

This is useful for sorting documents with certain tags such as car or todo prior to consumption. These folders won’t be deleted.

PAPERLESS_CONSUMER_RECURSIVE must be enabled for this to work.

Defaults to false.


Enables the scanning and page separation based on detected barcodes. This allows for scanning and adding multiple documents per uploaded file, which are separated by one or multiple barcode pages.

For ease of use, it is suggested to use a standardized separation page, e.g. here.

If no barcodes are detected in the uploaded file, no page separation will happen.

The original document will be removed and the separated pages will be saved as pdf.

Defaults to false.


Whether TIFF image files should be scanned for barcodes. This will automatically convert any TIFF image(s) to pdfs for later processing. This only has an effect, if PAPERLESS_CONSUMER_ENABLE_BARCODES has been enabled.

Defaults to false.


Defines the string to be detected as a separator barcode. If paperless is used with the PATCH-T separator pages, users shouldn’t change this.

Defaults to “PATCHT”


On smaller systems, or even in the case of Very Large Documents, the consumer may explode, complaining about how it’s “unable to extend pixel cache”. In such cases, try setting this to a reasonably low value, like 32. The default is to use whatever is necessary to do everything without writing to disk, and units are in megabytes.

For more information on how to use this value, you should search the web for “MAGICK_MEMORY_LIMIT”.

Defaults to 0, which disables the limit.


Similar to the memory limit, if you’ve got a small system and your OS mounts /tmp as tmpfs, you should set this to a path that’s on a physical disk, like /home/your_user/tmp or something. ImageMagick will use this as scratch space when crunching through very large documents.

For more information on how to use this value, you should search the web for “MAGICK_TMPDIR”.

Default is none, which disables the temporary directory.


Use optipng to optimize thumbnails. This usually reduces the size of thumbnails by about 20%, but uses considerable compute time during consumption.

Defaults to true.


After a document is consumed, Paperless can trigger an arbitrary script if you like. This script will be passed a number of arguments for you to work with. For more information, take a look at Post-consumption script.

The default is blank, which means nothing will be executed.


Paperless will check the document text for document date information. Use this setting to enable checking the document filename for date information. The date order can be set to any option as specified in The filename will be checked first, and if nothing is found, the document text will be checked as normal.

Defaults to none, which disables this feature.


Paperless creates thumbnails for plain text files by rendering the content of the file on an image and uses a predefined font for that. This font can be changed here.

Note that this won’t have any effect on already generated thumbnails.

Defaults to /usr/share/fonts/liberation/LiberationSerif-Regular.ttf.


Paperless parses a documents creation date from filename and file content. You may specify a comma separated list of dates that should be ignored during this process. This is useful for special dates (like date of birth) that appear in documents regularly but are very unlikely to be the documents creation date.

You may specify dates in a multitude of formats supported by dateparser (see but as the dates need to be comma separated, the options are limited. Example: “2020-12-02,22.04.1999”

Defaults to an empty string to not ignore any dates.


Paperless will try to determine the document creation date from its contents. Specify the date format Paperless should expect to see within your documents.

This option defaults to DMY which translates to day first, month second, and year last order. Characters D, M, or Y can be shuffled to meet the required order.


By default, paperless ignores certain files and folders in the consumption directory, such as system files created by the Mac OS.

This can be adjusted by configuring a custom json array with patterns to exclude.

Defaults to [".DS_STORE/*", "._*", ".stfolder/*", ".stversions/*", ".localized/*", "desktop.ini"].


There are a few external software packages that Paperless expects to find on your system when it starts up. Unless you’ve done something creative with their installation, you probably won’t need to edit any of these. However, if you’ve installed these programs somewhere where simply typing the name of the program doesn’t automatically execute it (ie. the program isn’t in your $PATH), then you’ll need to specify the literal path for that program.


Defaults to “/usr/bin/convert”.


Defaults to “/usr/bin/gs”.


Defaults to “/usr/bin/optipng”.

Docker-specific options

These options don’t have any effect in paperless.conf. These options adjust the behavior of the docker container. Configure these in docker-compose.env.


The number of worker processes the webserver should spawn. More worker processes usually result in the front end to load data much quicker. However, each worker process also loads the entire application into memory separately, so increasing this value will increase RAM usage.

Consider configuring this to 1 on low power devices with limited amount of RAM.

Defaults to 2.


The port number the webserver will listen on inside the container. There are special setups where you may need this to avoid collisions with other services (like using podman with multiple containers in one pod).

Don’t change this when using Docker. To change the port the webserver is reachable outside of the container, instead refer to the “ports” key in docker-compose.yml.

Defaults to 8000.


The ID of the paperless user in the container. Set this to your actual user ID on the host system, which you can get by executing

$ id -u

Paperless will change ownership on its folders to this user, so you need to get this right in order to be able to write to the consumption directory.

Defaults to 1000.


The ID of the paperless Group in the container. Set this to your actual group ID on the host system, which you can get by executing

$ id -g

Paperless will change ownership on its folders to this group, so you need to get this right in order to be able to write to the consumption directory.

Defaults to 1000.


Additional OCR languages to install. By default, paperless comes with English, German, Italian, Spanish and French. If your language is not in this list, install additional languages with this configuration option:


To actually use these languages, also set the default OCR language of paperless:


Defaults to none, which does not install any additional languages.

Update Checking


Enable (or disable) the automatic check for available updates. This feature is disabled by default but if it is not explicitly set Paperless-ngx will show a message about this.

If enabled, the feature works by pinging the the Github API for the latest release e.g. to determine whether a new version is available.

Actual updating of the app must still be performed manually.

Note that for users of thirdy-party containers e.g. this notification may be ‘ahead’ of a new release from the third-party maintainers.

In either case, no tracking data is collected by the app in any way.

Defaults to none, which disables the feature.