glot-run provides a http api for running code inside docker containers. The api is described here.

The download above is a standard erlang release that includes a start script. To start glot-run in the foreground type: glot/bin/glot foreground.

glot-run takes it's configuration from environment variables. All vars needs to be set, no default values are provided.

An api token is required to run code. Users can be created with the /admin/users endpoint. See the api docs for more details.

Languages can be added with the /admin/languages endpoint. A language has a name, version and the name of a docker image that will be used when running code for the given language/version. See the api docs for more details.

When a run request is posted to glot-run it will create a new temporary container from the image that handles the given language/version. The container is required to listen for a json payload on stdin and must write the run result to stdout as a json object containing the properties: stdout, stderr and error. An application that does this is glot-code-runner. Example images can be found here.

The payload {"files": [{"name": "main.py", "content": "print(42)"}]} posted to /languages/python/latest will result in this payload being sent to the container: {"language": "python", "files": [{"name": "main.py", "content": "print(42)"}]}. A successful run should yield this response from the container: {"stdout": "42\n", "stderr": "", "error": ""}.

By default the docker api is only available through a unix socket and it needs to be configured to listen on a tcp socket. On ubuntu this can be done by adding the line DOCKER_OPTS="-H tcp://0.0.0.0:2375 -H unix:///var/run/docker.sock" to /etc/default/docker.