01 / Audio
FOSSASIA Engineering Journey · GSoC 2026 chapter
VoxBento
Real-Time Speech Translation
and Multilingual Interpretation
.webp)
Verified activity
The contribution ledger
Build-time GitHub data makes this an auditable record, not a counter with a good imagination.
709Pull Requests
333Issues
18Repositories
645Merged PRs
The system
From spoken signal to shared understanding
Live audio path
Interpreter browserWebRTC / WHIPMediaMTXWebRTC / WHEPListener browser
Direct media path: Python coordinates it but does not process audio bytes.Interpretation pipeline
RTSP / FFmpegSpeech-to-Textfinal segmentsTranslationText-to-Speech
Captions, translations, and synthesized output are distributed by WebSocket.FastAPI · async coordinationWebSocket booth statePostgreSQL / SQLiteJitsi floor monitoringDocker · Linux · Nginx
Program arc
An extended 22-week work product
- Mar 03–09FOSSASIA SummitBangkok
- May 01–24Community BondingGSoC
- May 25–Aug 23Coding PeriodWeeks 01–13
- Jul 21–26WikimaniaParis
- Jul 27–Aug 05FOSSASIA CommunityGermany
- Aug 24–Nov 02Extended CodingWeeks 14–22
- Nov 02Work Product DueFinal report
Field notes
Latest updates
02 / Access
Speech becomes multilingual output.
RTSP and FFmpeg feed transcription; final segments move through translation, TTS, and live WebSocket delivery.03 / Upstream
FOSSASIA work stays inspectable.
645 merged PRs are tracked in the build-time FOSSASIA audit trail.