MiniHuey provides a very lightweight huey-like API that may be useful for certain classes of applications. Unlike Huey, the MiniHuey consumer runs inside a greenlet in your main application process. This means there is no separate consumer process to run, not is there any persistence for the enqueued/scheduled tasks; whenever a task is enqueued or is scheduled to run, a new greenlet is spawned to execute the task.

Usage and task declaration:

class MiniHuey([name='huey'[, interval=1[, pool_size=None]]])
  • name (str) – Name given to this huey instance.
  • interval (int) – How frequently to check for scheduled tasks (seconds).
  • pool_size (int) – Limit number of concurrent tasks to given size.
from huey import crontab
from huey.contrib.minimal import MiniHuey

huey = MiniHuey()

def fetch_url(url):
    return urllib2.urlopen(url).read()

@huey.task(crontab(minute='0', hour='4'))
def run_backup():


There is not a separate decorator for periodic, or crontab, tasks. Just use MiniHuey.task() and pass in a validation function.

When your application starts, be sure to start the MiniHuey scheduler:

from gevent import monkey; monkey.patch_all()

huey.start()  # Kicks off scheduler in a new greenlet.
start_wsgi_server()  # Or whatever your main application is doing...


Tasks enqueued manually for immediate execution will be run regardless of whether the scheduler is running. If you want to be able to schedule tasks in the future or run periodic tasks, you will need to call start().

Calling tasks and getting results works about the same as regular huey:

async_result = fetch_url('')
html = async_result.get()  # Block until task is executed.

# Fetch the Yahoo! page in 30 seconds.
async_result = fetch_url.schedule(args=('',),
html = async_result.get()  # Blocks for ~30s.