TTS, Translations & Image Management
Force-regenerate TTS for a text. Replaces old cached audio with fresh high-quality output.
Click Refresh to load stats.
Check whether a text is cached and measure DB + file-system lookup times. No audio is generated.
Runs a series of timed API calls and shows per-test latency. TTS generation will produce actual audio (cached afterwards).
Time a /tts/ request end-to-end, including queue wait. Use different texts to force fresh generation.
Times the slow endpoint specifically. Unique text combinations produce fresh generations (no cache). Set length constraints to measure the extra ffmpeg processing time.
Calls the OpenAI API directly with no queue, no cache, no database, no ffmpeg. Use this to determine whether the slowness is in your infrastructure or in OpenAI's API. Compare the time here vs. the Advanced TTS Timer above.