daily-v2 - Build Daily voice and multimodal agent pipelines
Provides workflows and references for building Daily or Pipecat voice and multimodal AI pipelines with service orchestration, function calling, and provenance tracking.
Tags
Updated: 2026-10-16Capabilities
What this skill does
- Create transport connections
- Initialize speech services
- Configure LLM context
- Assemble processing pipelines
- Implement function calling
- Handle interruptions and turn-taking
- Manage conversation context
- Deploy Pipecat Cloud agents
- Track upstream provenance
Inputs
- Application requirements
- Transport configuration
- AI service credentials
- Function schemas
- Copied upstream files
- Provenance metadata
Outputs
- Voice or multimodal pipeline
- Function call results
- Conversation context updates
- Deployment status
- Session and call metrics
- Provenance review notes
Requirements
- Daily or Pipecat-compatible transport
- STT, LLM, and TTS services
- VAD analyzer for turn detection
- Access to copied repository files
- Cloud deployment credentials when deploying
