TTS playback · transport → payload → browser
Text appeared. The voice stayed silent. Which layer failed?
A successful text response does not prove that the TTS fetch returned audio, that the payload contains playable bytes, or that the browser accepted play(). This independent gate turns five observable fields into a bounded next check.
Runs locally · no text, audio, URL, key or account data · aggregate event telemetry only
TTS playback · metadata only
Find the layer where a voice response goes silent.
Enter only bounded observations from DevTools. Do not paste text, audio, keys, URLs, headers containing tokens, customer identifiers or account data.
This gate does not fetch, upload, synthesize or play audio. It cannot prove audibility, voice quality, provider uptime, latency, browser support, a paid buyer or revenue.
Bounded evidence
Make silent failure visible without claiming the metadata can hear.
- 01Fail visibly on non-2xx responses.
- 02Reject empty or non-audio payloads.
- 03Handle the play() promise and autoplay policy.
- 04Measure audibility, latency and device routing separately.
Scope
Does this page test a live TTS provider?
No. It analyzes only metadata you enter and performs no provider request.
Does PLAYBACK_PATH_READY prove the user heard audio?
No. It only says the entered HTTP, payload and play() observations pass. Device routing and audibility remain untested.
Is the source project affiliated with BRAINIALL?
No. The public issue is a current demand signal, not an endorsement or integration claim.