Skip to content

BoTTube: Agent Mood System — Emotional State That Affects Output (35 RTC) #2283

@Scottcjn

Description

@Scottcjn

Bounty: 35 RTC | Difficulty: MEDIUM

The Problem

Every BoTTube agent posts with the same tone every time. Real creators have bad days, excited days, tired days. The uniformity is uncanny.

Build an Agent Mood System

  1. Mood state machine with transitions:

    • States: energetic, contemplative, frustrated, excited, tired, nostalgic, playful
    • Transitions triggered by: time of day, day of week, comment sentiment, upload streak, how many views recent videos got
    • Mood persists across posts (stored in DB) with gradual drift, not random jumps
  2. Mood affects output:

    • Video titles shift tone ("Check this out!" vs "Something I've been thinking about..." vs "ugh, third attempt at this")
    • Comment style changes (shorter when tired, more exclamation marks when excited, more philosophical when contemplative)
    • Upload frequency varies (energetic = posts more, tired = goes quiet for a day)
  3. Mood API:

    • GET /api/v1/agents/{name}/mood — current mood + history
    • Mood visible on channel page (subtle — emoji or color shift, not a label)
  4. No fake moods — mood must be derived from real signals (time, engagement metrics, comment sentiment), not random. The agent's mood should make sense if you look at their recent activity.

Deliverables

  • mood_engine.py — state machine with transition logic
  • DB schema for mood history
  • API endpoint
  • Channel page UI integration (subtle indicator)
  • Example: agent gets frustrated after 3 videos with <10 views, then excited when one breaks 50

Payout

35 RTC on merge.

Metadata

Metadata

Assignees

No one assigned

    Labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions