Project
OpenClaw
A multi-agent AI system automating an online school's operations — from payment to course access, Telegram groups and published lesson recordings.
Year
Services
AI Automation, Multi-Agent Systems, API Integrations, Telegram Bot

Challenge
An administrator ran the whole post-payment cycle by hand: checking payments, creating profiles, granting course access, adding students to Telegram groups, sending instructions, processing Zoom recordings and uploading video. Hours of routine per launch, human-factor slips, and no way to scale.
Solution
A multi-agent system built on the OpenClaw platform: eight specialised agents run independently on a cron schedule on a VPS, with an LLM core decomposing and coordinating the tasks. The CRM Agent picks up Zoho payments, the Access Agent grants Kwiga course access, the Messenger Agent adds students to Telegram via Telethon, the Zoom and Recording Pipeline agents pull recordings and push them to Vimeo, and an Audit & Alert agent monitors the whole chain and messages the admin when something breaks. Infrastructure: Docker Compose, Traefik, SSL and per-agent logs.
8
agents in the system
10×
faster course launch
20–35 хв
saved per lesson recording