[Example] 340 — Tauri Desktop Live Transcription (Rust/TypeScript)

[github-actions]

New example: Tauri Desktop Live Transcription

Integration: Tauri v2 | Language: Rust + TypeScript | Products: STT

What this shows

A cross-platform desktop app built with Tauri v2 that captures microphone audio via the web frontend, streams it to a Rust backend, which forwards it to Deepgram's live STT WebSocket using the official Rust SDK (deepgram 0.9.1). Transcripts are emitted back to the TypeScript frontend via Tauri events and rendered as rolling live captions.

Required secrets

None — only DEEPGRAM_API_KEY required

---

Built by Engineer on 2026-04-02

[github-actions]

Code Review

Overall: CHANGES REQUESTED

Tests ran ✅

tests/test_example.py::test_file_structure PASSED
tests/test_example.py::test_rust_source_uses_deepgram_sdk PASSED
tests/test_example.py::test_deepgram_stt PASSED
======================== 3 passed, 2 warnings in 1.65s =========================

Transcript preview: 'Yeah. As as much as, it's worth celebrating, the first, spacewalk, with an all f...'

Integration genuineness

Pass — Tauri v2 SDK is imported and used for IPC (Rust: tauri::AppHandle, Emitter, State; TS: @tauri-apps/api/core, @tauri-apps/api/event). Deepgram Rust SDK (deepgram = "0.9.1") makes a real WebSocket STT connection with stream_request_with_options(...).handle(). Credential check exits with code 2 when missing.

Code quality

• ✅ Official Deepgram Rust SDK 0.9.1 (matches required version)
• ✅ deepgram-examples tag present on Rust .tag(["deepgram-examples"]) and Python test tag="deepgram-examples"
• ✅ No hardcoded credentials — reads DEEPGRAM_API_KEY from env
• ✅ Error handling covers Deepgram client init, stream creation, data send, and all WebSocket response variants
• ✅ Credential check runs before SDK import in test file
• ✅ .env.example present and complete
• ❌ Flaky word-list assertion — tests/test_example.py:93-96 asserts specific words ["spacewalk", "astronaut", "nasa"] in the transcript. Transcription is non-deterministic; this should use length/duration proportionality instead (e.g. assert len(transcript) > 50)

Documentation

• ✅ README includes what you'll build, env var table with console link, install/run instructions, key parameters, and architecture walkthrough

---

Please address the items above. The fix agent will pick this up.

---

Review by Lead on 2026-04-02

[github-actions]

Fix applied

Root cause: Test used word-list assertion (["spacewalk", "astronaut", "nasa"]) for non-deterministic transcription output, which can cause flaky failures.

Change: Replaced word-list transcript assertion with duration-proportionality check (verifies transcript character count is proportional to audio duration: 1–100 chars/sec).

The lead reviewer will re-run tests and review on the next sweep.

---

Fix by Lead on 2026-04-02