WeCron: a WeChat reminder service on Django 1.9.1 and python:2.7
Scheduled reminders on WeChat - Cron on WeChat
At a glance
- What is it?
- polyrabbit/WeCron delivers scheduled reminders through a WeChat public account from a Django application pinned to Python 2 era dependencies, whose container image installs an SSH daemon alongside the app and whose scheduler recovers missed reminders from a single custom management command before uwsgi starts.
- Who is it for?
- WeCron answers a narrow question that most scheduling tools skip, which is whether a reminder still arrives after the machine restarts, and the answer lives in one line of the Makefile rather than in a design document. That is worth reading whether or not you deploy it.
- Can I use it commercially?
- Yes, with conditions. GPL-3.0 is a copyleft licence: if you distribute software that includes it, you must release that software's source code under the same licence. Running it internally without distributing it does not trigger that obligation.
- Is it still maintained?
- Yes. The repository received new commits within the last day.
- What is it written in?
- Mainly JavaScript, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on November 15, 2026, and from our analysis. They are not legal advice.
Star history
- polyrabbit/WeCron694 ★ · +3
| Repository | 2026-09-17 | 2026-11-15 |
|---|---|---|
| polyrabbit/WeCron | 691 | 694 |
3 Stars in the last 59 days, measured from the daily GitHub snapshot. Compare with other projects
Editorial analysis
The quick start changes directory and then uses root-level paths
The local setup instructions are four steps, and two of them disagree with each other. Step two is `cd WeCron` followed by `pip install -r requirements.txt`. Step three is `python WeCron/manage.py migrate`, and step four is `python WeCron/manage.py runserver`. Both of those path prefixes only make sense from the directory above `WeCron/`, and `requirements.txt` is a top-level entry in the repository listing alongside `Dockerfile`, `Makefile`, `Procfile`, `fly.toml`, `manifest.yml`, and `deploy/`. In other words the install step expects you to be inside the app package while the run steps expect you to be outside it. That is a small thing, but it is the difference between a first run that works and one that stops on a missing file, and there is no other written setup path to fall back on, since the README is forty-odd words of Chinese plus these commands.
Every dependency is pinned to a Python 2 era version
`requirements.txt` is twenty-one lines and almost every one carries an exact pin. `Django==1.9.1`, `djangorestframework==3.6.4`, `uWSGI==2.0.19`, `requests==2.8.1`, `psycopg2==2.8.6`, `APScheduler==3.0.6`, and `wechatpy==1.3.1`, the WeChat platform SDK contributed by a third party the project credits in its own notes. `Pillow==6.2.2`, `jieba==0.42.1` for Chinese word segmentation, `pydub==0.25.1` for audio, `shove==0.6.6` for pushing files to object storage, `grip==4.6.2` for rendering GitHub-flavoured Markdown locally, and `httmock` for HTTP mocking in tests. The two that deserve attention are `pycrypto==2.6.1`, which has been unmaintained for years and carries published advisories, and the pair of very small helpers `ago==0.0.6` and `tomorrow==0.2.4`. Neither is a large surface, but both are exactly pinned, so nothing moves without editing the file.
missing_reschedule is how a missed reminder comes back
This is the mechanism worth understanding, and it is one target in the Makefile. The production target reads `run-in-prod: clean syncdb collectstatic reschedule`, and the last of those runs `python WeCron/manage.py missing_reschedule` before uwsgi is exec'd. So every boot does three things in order: delete stray `.pyc` files and the collected static directory, apply migrations, and then run a custom management command whose name says exactly what it does, find reminders that have no scheduled job and re-register them. For a service whose entire value is that a message arrives at a particular minute, that recovery step is the difference between a cron that survives a deploy and one that silently stops reminding people. The scheduling design itself is written up separately on the author's blog, linked from the README, and the runtime scheduler underneath is APScheduler.
The container runs sshd as root next to the web process
The Dockerfile starts from `python:2.7` and carries a `MAINTAINER` line, which is a directive modern builders warn about and ignore. It sets `VCAP_APP_PORT=8000`, the Cloud Foundry convention, and adds a health check that curls `/?healthy` on that port. Alongside the application it installs `ffmpeg` and `libavcodec-extra`, which lines up with the `pydub` dependency and suggests the service handles voice messages as well as text. Then it installs an OpenSSH server, copying `vim` and `lsof` in with it, rewriting `sshd_config` so the daemon listens on port 2333 with `PasswordAuthentication no`, creating `/root/.ssh` with mode 700, and finally running `service ssh start && make run-in-prod` as the container command. For a reminder bot, that is a shell into a production container as root, and the image is where you would have to change that.
Four deployment targets are configured side by side
The repository root carries `fly.toml`, a `Procfile`, a `manifest.yml`, a `Dockerfile`, and a `deploy/` directory, which is four different platforms configured at once. The `manifest.yml` and the `VCAP_APP_PORT` variable inside the Dockerfile are both Cloud Foundry conventions. `fly.toml` is Fly.io. The `Procfile` is the Heroku-era format. And `make release` runs `test` and then `ansible-playbook deploy/playbook.yml -v`, so the `deploy/` directory holds a real Ansible path as well. The Makefile itself has accumulated drift in the same places: `run-in-prod` keeps the `uwsgi --ini=...j2` line commented out next to the live one, `run-uwsgi-test` hardcodes `--processes=4` while the variable `UWSGI_WORKERS ?= 1` declared at the top is never referenced by any recipe, and `docs` renders `raas.md` through grip with the title translated on the command line.
The database is PostgreSQL, bootstrapped by two psql commands
There is one supported store and it is set up by hand rather than by a script:
psql -c 'CREATE DATABASE "wecron" WITH OWNER "postgres" TEMPLATE template0 ENCODING="UTF8" CONNECTION LIMIT=-1;'
psql -c 'GRANT ALL PRIVILEGES ON DATABASE "wecron" to "postgres";'The bootstrap creates a database called `wecron` with owner `postgres`, `TEMPLATE template0`, `ENCODING="UTF8"`, and `CONNECTION LIMIT=-1`, then grants all privileges on it. Only then does `python WeCron/manage.py migrate` create the tables. Two details in that first command matter for a reminder service: `template0` is chosen so the encoding is set rather than inherited, and an unlimited connection limit is what lets a long-lived scheduler process hold its jobs. The Makefile's `syncdb` target wraps the migration step as `migrate --noinput`, which is what the container start path uses. `psycopg2==2.8.6` is the only database driver in the requirements, so there is no alternative store behind a setting.
What the service is, and where the write-up lives
The scope is narrower than the repository layout suggests. It is scheduled reminders delivered through WeChat, described in the repository as a cron for WeChat, and the delivery channel is a public account you follow by scanning a QR code, with a live demo linked from the README. There is no web UI in the documentation, no REST surface beyond `djangorestframework` sitting in the requirements, and no billing or multi-tenancy story. The author's own account of how the scheduling works is hosted on a personal blog under a post about how WeCron schedules tasks, and that post, together with the `missing_reschedule` target, is the most reusable thing here. On maintenance status, the concrete facts are these: the project says it began in 2015, the last push to `master` is dated 2026-03-16, and the repository has no GitHub releases, so there is no version history to pin against or to diff.
Editorial conclusion
WeCron answers a narrow question that most scheduling tools skip, which is whether a reminder still arrives after the machine restarts, and the answer lives in one line of the Makefile rather than in a design document. That is worth reading whether or not you deploy it. Whether you deploy it is a separate question, because the dependency set is from the Python 2 era, `pycrypto` is pinned at a version with known advisories against it, the container runs an SSH daemon as root next to a web process, and the database bootstrap is PostgreSQL only. The last push to `master` is 2026-03-16 and the repository has no tagged releases, so there is no version history to pin against. If your use case is a personal reminder bot and you already have a WeChat public account, the design write-up on how the scheduling works is the most transferable part here; if it is anything customer-facing, treat the dependency and operational footprint as the problem to solve before the reminders are.
Frequently asked questions
What is WeCron and how does it deliver reminders?
It is a scheduled reminder service that delivers messages through a WeChat public account, described in the repository as a cron for WeChat. You follow the account by scanning a QR code, and reminders arrive as messages. The WeChat platform SDK it uses is wechatpy, contributed by a third party the project credits.
What does WeCron need in order to run locally?
PostgreSQL and a Python 2 era environment. Two `psql -c` commands create a `wecron` database owned by `postgres` with UTF8 encoding and grant privileges, then `python WeCron/manage.py migrate` builds the tables and `python WeCron/manage.py runserver` starts the server. Dependencies come from the root `requirements.txt`, which pins Django at 1.9.1.
How does WeCron recover reminders after a restart?
The production target in the Makefile chains `clean`, `syncdb`, `collectstatic`, and `reschedule` before exec'ing uwsgi. The reschedule step runs the custom management command `missing_reschedule`, which re-registers reminders that have no pending job before the server starts accepting traffic.
Is WeCron still being worked on?
The last push to `master` is dated 2026-03-16, and the repository has no GitHub releases, so there is no tagged history. The README says the project began in 2015 and states an intention to keep maintaining it, but no commits have landed since that date.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/polyrabbit-wecron)Star history chart for your README
[](https://hysenlabs.com/projects/polyrabbit-wecron)Was this useful?