One Pipecat pipeline, six voices: classifier-fronted handover between AI characters, one-shotted by Opus
alexcovo_eth · x · 2026-11-01
JonPTaylor shared an open-source Pipecat demo (pipecat-one-pipeline-two-characters, 47 stars): ONE bot pipeline with a classifier front-running the LLM to determine which character the user is addressing, then Pipecat orchestrates handovers between character workers — an idea he says Opus 5.5 one-shotted.
The demo, "one room, six voices," seats six characters (botanist, chef, musician, sea captain, engineer, historian) around a kitchen table, each with their own Cartesia voice, Daily audio track, own LLM and on-screen color, all hearing everything. Quick questions get near-simultaneous answers starting moments apart; proper questions are answered in turn; addressing one character means only they reply — a room of people, not a queue.
More from coding & agent
- Dev shares Grok Bot team-building guide: few focused bots that write their own skills — CurieuxExplorer · 2026-11-01
- 747 AI-agent ideas reveal a scaling law for idea descent: +17% per doubling — DimitrisPapail · 2026-11-01
- Magnific Move + Claude Opus: Halloween game animated in one chat — petewoodbridge · 2026-11-01
- AI-Driven Frame Rate Optimization Cuts Render Time by 9%-15% Across Six Scenes — AIandDesign · 2026-11-01
- Parcha founder: You may not need an agent after all — decision models vs workflows — blaizedsouza · 2026-11-01
- Claude Code agents burned 9.45M tokens a day reading their own notes — WWEtrump · 2026-11-01