Work

We run this stack ourselves

Both systems below run our own venues. That is deliberate: we would rather show you something carrying real money and real staff every day than a pilot we talked a client into.

Noisy Chili Tap House

Replacing a POS licence with a Telegram mini app

A tap house running on Loyverse wanted off the per-till licence without going back to paper. We moved the catalogue and rebuilt the till as a Telegram mini app.

In-house deployment — Noisy Chili and Riel Brewing are our own venues.

Point of sale mini app interface

The problem

Monthly per-till cost, a tablet that had to survive a bar, and no way to check takings without opening a back office on a laptop.

What we did

  • Exported and cleaned the Loyverse catalogue: 185 products, 10 categories, 5 pricing rules
  • Imported to Postgres in a single verified pass with row-count reconciliation
  • Built the till as a Telegram mini app against an action-routed API
  • Added a separate admin API with role checks resolved from the Telegram account
  • Scheduled a nightly sales summary to the manager's chat at 23:30

Where it landed

  • Every staff phone is a till; no dedicated hardware to buy or break
  • Takings visible on demand, and pushed nightly without anyone asking
  • Admin actions gated server-side, so staff access does not depend on UI hiding
  • No per-seat licence
Noisy Chili Tap House

Eight platforms, one approval step

Event promotion was not happening because posting to five apps by hand is a chore nobody picks up. We built a queue with a review step and let a schedule do the rest.

In-house deployment.

Social post approval queue interface

The problem

Events went unpromoted. Fully automatic posting was not acceptable — the venue's voice matters and an unreviewed post is a liability.

What we did

  • Queue held in a Data Table, one row per post, with an explicit approved flag
  • Telegram mini app to preview, edit, approve or deny each post
  • Scheduler sweeps every 15 minutes and publishes only approved, due posts
  • Fan-out to eight destinations, each called and reported separately
  • AI flyer builder generates artwork when a post has no photo

Where it landed

  • Nothing publishes without a human approving it
  • Posts with no photo still go out, with generated artwork
  • Per-platform failures surface immediately in Telegram instead of silently
  • Review takes under a minute from a phone
Also in production

Smaller systems, same approach

Order intake and invoicing

Customers message an order in plain language. It is parsed against the live catalogue, matched to a customer record, filed, and confirmed. A real PDF invoice is generated and delivered to their chat. Open orders are chased with a 9am digest.

Customer service bot

A Telegram front desk that has been answering for the Automaton Labs shop since August 2025 — catalogue, orders and profile, with admin actions behind an authenticated boundary.

Staff operations

Opening and closing checklists signed off from a mini app, shift reminders to the staff group, flagged issues escalated, and a Monday digest of the week's sign-offs.

Environmental monitoring

Continuous temperature and humidity logging with threshold alarms that escalate to a person. Built for cold rooms and fermentation, where a four-hour excursion is a write-off.

Tell us what keeps going wrong

A conversation costs nothing and often ends with us saying automation is not worth it for a given job. That is a useful answer too.