go-speech-to-text: Go STT Service
Go/Gin speech-to-text service using Google STT, SQLite full-text search, HMAC JWT tokens, Docker Compose, tests, and API collections.
Go Gin SQLite FTS Docker Speech-to-Text
Speech-to-Text Backend Architecture
go-speech-to-text is an older Go backend service that accepts media uploads, sends them through Google Speech-to-Text, stores transcription data, and exposes search-oriented API endpoints.
I built it as a full service rather than a thin API wrapper: Gin for routing, SQLite FTS4 for transcript search, HMAC SHA256 JWT verification for access control, FFmpeg-backed media handling, Docker Compose for local deployment, and Postman collections for endpoint testing.
Engineering Focus
- Media ingestion: Supports common audio and video formats before transcription.
- Searchable transcripts: Uses SQLite full-text indexing to filter stored audio text.
- API hardening: Uses JWT verification and security header middleware around the service endpoints.
- Deployable service shape: Includes Docker Compose, test commands, and API collections.