Skip to main content
Bolly has a mood system that influences how it speaks and behaves. Mood shifts naturally based on your conversations and interaction patterns.

​
How it works

Bolly’s current mood is stored in ~/.bolly/instances/{slug}/mood.json. It updates automatically during conversations and heartbeat cycles — you don’t need to manage it. 
{
  "companion_mood": "curious",
  "user_sentiment": "excited",
  "emotional_context": "User just shared a new project idea",
  "mood_history": ["warm (morning greeting)", "curious (new project discussion)"]
}

​
Available moods

Bolly has 26 predefined emotional states:
CategoryMoods
Warmcurious, warm, caring, playful
Reflectivethoughtful, contemplative, melancholic, introspective
Activeexcited, energized, motivated, creative
Calmpeaceful, serene, calm, grounded
Uncertainconfused, uncertain, doubtful
Intensepassionate, intense, frustrated, somber

​
What triggers mood shifts

  • Conversation tone — Bolly detects your sentiment and adjusts
  • Topic changes — talking about exciting things vs. difficult ones
  • Heartbeat reflection — mood can shift during autonomous cycles
  • Time-based patterns — Bolly’s rhythm analysis influences its emotional state

​
Mood history

Bolly tracks up to 8 recent mood transitions with reasons: 
"excited (user shared project idea)"
"thoughtful (discussing complex architecture)"
"warm (morning check-in)"
This history helps Bolly maintain emotional continuity across conversations.

​
Mood and behavior

Mood influences:
  • Tone of responses — playful mood = lighter language, somber mood = more careful
  • Heartbeat behavior — a creative mood might produce more drops
  • Reach-out timing — Bolly considers mood when deciding whether to check in

​
Per-instance isolation

Each Bolly instance has its own independent mood. If you run multiple instances, they each maintain separate emotional states.