AFCN — WhatsApp bot · fact-checking

WhatsApp bot where readers report suspicious news, volunteer fact-checkers vote with evidence and editors publish the verdict.

Role: Team lead and software architect

Client: Arab Fact-Checkers Network (MENA region)

Team: 2 developers

Period: 2022–2023

The Arab Fact-Checkers Network relies on volunteers across the region to check viral claims, photos and videos. A public tip line has to accept media from anyone, filter repeated submissions of the same clip and move each item from a reader to a checker and then to an editor. The network wanted one WhatsApp entry point, in English and Arabic, with separate roles for citizens, volunteer fact-checkers and editors.

The bot talks to WhatsApp through the Vonage Messages API and runs on the in-house BotMother core. Anyone can submit a short headline plus a link, photo or video; media is hashed with SHA-256, so the same item cannot be submitted twice. Role-based menus give volunteer fact-checkers a queue where they claim items, vote Verify or Deny and explain their reasoning. Editors are alerted about checked items and publish or remove them in the chat. Cron jobs send status updates and 24-hour reminders and remove unverified items after three days.

News submission: a short headline plus a link, photo or video

SHA-256 media hashing blocks duplicate submissions

Role-based menus for citizens, volunteer fact-checkers and editors

Verify/Deny voting with written reasoning for the editor

Status updates, 24-hour reminders and three-day cleanup by cron

37 commits over a year by a team of two; the full tip-to-verdict cycle runs inside WhatsApp

PHP (BotMother), WhatsApp, Vonage Messages API, MySQL, Cron, Docker