Binary Release

[ Download from release page ]

The release page has pre-compiled binaries for Mac OS X, Windows, and several Linux distributions. Extract the tarball and run the binary inside with the --help flag to see usage instructions:

# Untar the release (available at

$ tar Jxf postgrest-[version]-[platform].tar.xz

# Try running it
$ ./postgrest --help

# You should see a usage help message


If you see a dialog box like this on Windows, it may be that the pg_config program is not in your system path.


It usually lives in C:\Program Files\PostgreSQL\<version>\bin. See this article about how to modify the system path.

PostgreSQL dependency

To use PostgREST you will need an underlying database (PostgreSQL version 9.5 or greater is required). You can use something like Amazon RDS but installing your own locally is cheaper and more convenient for development.

On Windows, PostgREST will fail to run unless the PostgreSQL binaries are on the system path. To test whether this is the case, run pg_config from the command line. You should see it output a list of paths.


The PostgREST server reads a configuration file as its only argument:

./postgrest /path/to/postgrest.conf

For a complete reference of the configuration file, see Configuration.

Running the Server

PostgREST outputs basic request logging to stdout. When running it in an SSH session you must detach it from stdout or it will be terminated when the session closes. The easiest technique is redirecting the output to a log file or to the syslog:

ssh [email protected] \
  'postgrest foo.conf </dev/null >/var/log/postgrest.log 2>&1 &'

# another option is to pipe the output into "logger -t postgrest"


You can get the official PostgREST Docker image with:

docker pull postgrest/postgrest

The image consults an internal /etc/postgrest.conf file. To customize this file you can either mount a replacement configuration file into the container, or use environment variables. The environment variables will be interpolated into the default config file.

These variables match the options shown in our Configuration section, except they are capitalized, have a PGRST_ prefix, and use underscores. To get a list of the available environment variables, run this:

docker inspect -f "{{.Config.Env}}" postgrest/postgrest

There are two ways to run the PostgREST container: with an existing external database, or through docker-compose.

Containerized PostgREST with native PostgreSQL

The first way to run PostgREST in Docker is to connect it to an existing native database on the host.

# Run the server
docker run --rm --net=host -p 3000:3000 \
  -e PGRST_DB_URI="postgres://[email protected]/postgres" \
  -e PGRST_DB_ANON_ROLE="postgres" \

The database connection string above is just an example. Adjust the role and password as necessary. You may need to edit PostgreSQL’s pg_hba.conf to grant the user local login access.


Docker on Mac does not support the --net=host flag. Instead you’ll need to create an IP address alias to the host. Requests for the IP address from inside the container are unable to resolve and fall back to resolution by the host.

sudo ifconfig lo0 alias

You should then use as the host in your database connection string. Also remember to include the IP address in the listen_address within postgresql.conf. For instance:

listen_addresses = 'localhost,'

You might also need to add a new IPv4 local connection within pg_hba.conf. For instance:

host    all             all               trust

Containerized PostgREST and db with docker-compose

To avoid having to install the database at all, you can run both it and the server in containers and link them together with docker-compose. Use this configuration:

# docker-compose.yml

version: '3'
    image: postgrest/postgrest
      - "3000:3000"
      - db:db
      PGRST_DB_URI: postgres://app_user:[email protected]:5432/app_db
      PGRST_DB_SCHEMA: public
      PGRST_DB_ANON_ROLE: app_user #In production this role should not be the same as the one used for the connection
      - db
    image: postgres
      - "5432:5432"
      POSTGRES_DB: app_db
      POSTGRES_USER: app_user
      POSTGRES_PASSWORD: password
  # Uncomment this if you want to persist the data.
  # volumes:
  #   - "./pgdata:/var/lib/postgresql/data"

Go into the directory where you saved this file and run docker-compose up. You will see the logs of both the database and PostgREST, and be able to access the latter on port 3000.

If you want to have a visual overview of your API in your browser you can add swagger-ui to your docker-compose.yml:

  image: swaggerapi/swagger-ui
    - "8080:8080"
    - "8080"
    API_URL: http://localhost:3000/

With this you can see the swagger-ui in your browser on port 8080.

Deploying to Heroku

Assuming your making modifications locally and then pushing to GitHub, it’s easy to deploy to Heroku.

  1. Create a new app on Heroku
  2. In Settings add the following buildpack
  3. Add the require Config Vars in Heroku (see for more details)
  4. Modify your postgrest.conf file as required to match your Config Vars in Heroku
  5. Create your Procfile and add ./env-to-config ./postgrest postgrest.conf
  6. Push your changes to GitHub
  7. Set Heroku to automatically deploy from Master and then manually deploy the branch for the first build

Build from Source


We discourage building and using PostgREST on Alpine Linux because of a reported GHC memory leak on that platform.

When a pre-built binary does not exist for your system you can build the project from source. You’ll also need to do this if you want to help with development. Stack makes it easy. It will install any necessary Haskell dependencies on your system.

  • Install Stack for your platform

  • Install Library Dependencies

    Operating System Dependencies
    Ubuntu/Debian libpq-dev, libgmp-dev
    CentOS/Fedora/Red Hat postgresql-devel, zlib-devel, gmp-devel
    BSD postgresql95-client
    OS X libpq, gmp
  • Build and install binary

    git clone
    cd postgrest
    # adjust local-bin-path to taste
    stack build --install-ghc --copy-bins --local-bin-path /usr/local/bin


If building fails and your system has less than 1GB of memory, try adding a swap file.

  • Check that the server is installed: postgrest --help.

PostgREST Test Suite

Creating the Test Database

To properly run postgrest tests one needs to create a database. To do so, use the test creation script create_test_database in the test/ folder.

The script expects the following parameters:

test/create_test_db connection_uri database_name [test_db_user] [test_db_user_password]

Use the connection URI to specify the user, password, host, and port. Do not provide the database in the connection URI. The PostgreSQL role you are using to connect must be capable of creating new databases.

The database_name is the name of the database that stack test will connect to. If the database of the same name already exists on the server, the script will first drop it and then re-create it.

Optionally, specify the database user stack test will use. The user will be given necessary permissions to reset the database after every test run.

If the user is not specified, the script will generate the role name postgrest_test_ suffixed by the chosen database name, and will generate a random password for it.

Optionally, if specifying an existing user to be used for the test connection, one can specify the password the user has.

The script will return the db uri to use in the tests–this uri corresponds to the db-uri parameter in the configuration file that one would use in production.

Generating the user and the password allows one to create the database and run the tests against any PostgreSQL server without any modifications to the server. (Such as allowing accounts without a password or setting up trust authentication, or requiring the server to be on the same localhost the tests are run from).

Running the Tests

To run the tests, one must supply the database uri in the environment variable POSTGREST_TEST_CONNECTION.

Typically, one would create the database and run the test in the same command line, using the postgres superuser:

POSTGREST_TEST_CONNECTION=$(test/create_test_db "postgres://postgres:[email protected]" test_db) stack test

For repeated runs on the same database, one should export the connection variable:

export POSTGREST_TEST_CONNECTION=$(test/create_test_db "postgres://postgres:[email protected]" test_db)
stack test
stack test

If the environment variable is empty or not specified, then the test runner will default to connection uri


This connection assumes the test server on the localhost:code: with the user postgrest_test without the password and the database of the same name.

Destroying the Database

The test database will remain after the test, together with four new roles created on the PostgreSQL server. To permanently erase the created database and the roles, run the script test/delete_test_database, using the same superuser role used for creating the database:

test/destroy_test_db connection_uri database_name

Testing with Docker

The ability to connect to non-local PostgreSQL simplifies the test setup. One elegant way of testing is to use a disposable PostgreSQL in docker.

For example, if local development is on a mac with Docker for Mac installed:

$ docker run --name db-scripting-test -e POSTGRES_PASSWORD=pwd -p 5434:5432 -d postgres
$ POSTGREST_TEST_CONNECTION=$(test/create_test_db "postgres://postgres:[email protected]:5434" test_db) stack test

Additionally, if one creates a docker container to run stack test (this is necessary on Mac OS Sierra with GHC below 8.0.1, where stack test fails), one can run PostgreSQL in a separate linked container, or use the locally installed PostgreSQL app.

Build the test container with test/Dockerfile.test:

$ docker build -t pgst-test - < test/Dockerfile.test
$ mkdir .stack-work-docker ~/.stack-linux

The first run of the test container will take a long time while the dependencies get cached. Creating the ~/.stack-linux folder and mapping it as a volume into the container ensures that we can run the container in disposable mode and not worry about subsequent runs being slow. .stack-work-docker is also mapped into the container and must be specified when using stack from Linux, not to interfere with the .stack-work for local development. (On Sierra, stack build works, while stack test fails with GHC 8.0.1).

Linked containers:

$ docker run --name pg -e POSTGRES_PASSWORD=pwd  -d postgres
$ docker run --rm -it -v `pwd`:`pwd` -v ~/.stack-linux:/root/.stack --link pg:pg -w="`pwd`" -v `pwd`/.stack-work-docker:`pwd`/.stack-work pgst-test bash -c "POSTGREST_TEST_CONNECTION=$(test/create_test_db "postgres://postgres:[email protected]" test_db) stack test"

Stack test in Docker for Mac, PostgreSQL app on mac:

$ host_ip=$(ifconfig en0 | grep 'inet ' | cut -f 2 -d' ')
$ export POSTGREST_TEST_CONNECTION=$(test/create_test_db "postgres://[email protected]$HOST" test_db)
$ docker run --rm -it -v `pwd`:`pwd` -v ~/.stack-linux:/root/.stack -v `pwd`/.stack-work-docker:`pwd`/.stack-work -e "HOST=$host_ip" -e "POSTGREST_TEST_CONNECTION=$POSTGREST_TEST_CONNECTION" -w="`pwd`" pgst-test bash -c "stack test"
$ test/destroy_test_db "postgres://[email protected]" test_db