Description

This PR implements a complete Deepgram piece integration that provides all 5 required actions for voice AI automation workflows. The integration includes speech-to-text transcription, audio summarization, text-to-speech conversion, and project management capabilities.

Changes Made

Created complete Deepgram piece structure: Implemented proper directory structure following src/actions/ and src/common/ patterns Added all required configuration files (package.json, project.json, tsconfig files) Added Deepgram piece to tsconfig.base.json paths

1 Implemented 5 core actions: Create Summary: Produces intelligent summaries of audio content with topic detection and sentiment analysis Transcribe Audio: High-accuracy speech-to-text conversion with speaker diarization Create Transcription (Callback): Asynchronous transcription processing with webhook support List Projects: Account and project management functionality Text to Speech: Natural-sounding speech generation with multiple voice models

  1. Added robust authentication: API Key authentication with real-time validation using /projects endpoint Comprehensive error handling with proper status code checking Clear setup instructions and user guidance

  2. Implemented proper file handling: Correct Buffer handling for audio input/output operations Integration with Activepieces files service for TTS output Support for multiple audio formats (MP3, WAV, FLAC, AAC, OGG, etc.)

  3. Added comprehensive error handling: Contextual error messages with troubleshooting guidance Proper HTTP status code handling and API error mapping Input validation with helpful user feedback

  4. Created HTTP client abstraction: Clean DeepgramClient class with reusable methods Generic typing for proper TypeScript support Support for different response types (JSON, ArrayBuffer)

Testing

Authentication validation tested with real API calls All property validations and type checking verified File upload/download operations tested Error scenarios handled properly TypeScript compilation successful

Related Issues

This implementation provides a complete solution for Deepgram voice AI automation, enabling users to build sophisticated audio processing workflows.

/fixes #7931 /claim #7931

Claim

Total prize pool $50
Total paid $0
Status Pending
Submitted June 03, 2025
Last updated June 03, 2025

Contributors

KU

Kunal Darekar

@Kunal-Darekar

100%

Sponsors

AC

Activepieces (YC S22)

@activepieces

$50