The auto-reply engine runs as background workers that continuously process tweets without manual intervention.

Architecture

GenX runs two background workers:

Campaign Poller (every 30s)

  • Scans all active campaigns
  • Checks if the campaign’s search interval has elapsed
  • Searches Twitter for matching tweets
  • Creates conversations from new matches
  • Triggers auto-generate or auto-reply based on campaign settings

Auto-Reply Worker (every 60s)

  • Finds all users with pending conversations
  • Groups conversations by campaign
  • Generates AI replies using campaign-specific tone and prompt
  • Posts replies using the assigned bot account
  • Updates campaign statistics
  • Sends email notifications

Processing Priority

  1. Campaign-level settings override global settings
  2. Bot account resolution: Campaign’s designated bot > random logged-in bot
  3. Proxy resolution: Account-assigned proxy > any active user proxy

Enabling Auto-Reply

Per-campaign

Toggle Auto-Reply when creating or editing a campaign. This enables full autopilot for that campaign only.

Per-campaign (generate only)

Toggle Auto-Generate to automatically create drafts without posting. You still approve manually.

Global

Enable Auto-Reply in Settings to process all pending conversations, including those not tied to a campaign.

Reply Quality Controls

Generated replies are automatically cleaned:
  • Hard limit of 220 characters (reserves space for @mention)
  • No URLs, hashtags, or markdown
  • Maximum 2 emojis
  • Sentence-aware truncation
  • Preamble removal (strips “Sure!”, “Here’s a reply:”, etc.)

Reply Tones

ToneStyle
Casual FriendlyFriendly and approachable
ProfessionalPolished and business-appropriate
Witty HumorousClever and engaging
Supportive EmpatheticUnderstanding and warm
InformativeHelpful and fact-driven
ProvocativeBold and thought-provoking