Voxta Service tests
Run every installed module once — text, voice, transcription, images, vision, memory, video and animation — and see what passes, how fast, and how much VRAM it took.
Advanced → Service Tests (administrators) runs a short test on every installed module, one service at a time, and reports what passed. Click Start Tests; Run Again once it finishes, or the play button on one card to run just that test.
When to use it
- Right after install to confirm everything's wired up.
- After an upgrade to verify nothing regressed.
- When something feels broken — it narrows the issue to one module quickly.
- Before buying a GPU or picking a model — each card shows the VRAM it needed.
Choosing what runs
- Service type — All, Text Gen, TTS, STT, Transcription, Image Gen, Vision, Memory, 3D Animations, Video Gen, Portrait Animation.
- Hosting — All, Local, Built-in, External, Online.
- The toggle on a card enables or disables that module. Disabled modules are skipped.
- Pause on failure stops at the first failed test: Continue to skip it, Retry, or Stop.
- Measure VRAM in isolation unloads every model before each test so each one measures only what it loads. Slower: every test pays for a cold load.
Modules that still need installing are listed but not tested.
Reading results
Each card shows passed, warning, failed or skipped, with:
- Time — the first run (including loading the model) and the second run (already loaded).
- VRAM — the most the test needed at once, and what stayed held once it was idle.
- Output — the audio, image, clip or animation it produced, to check by eye and ear.
- Error — why it failed. Usually points at the fix (bad API key, missing model file…).
The gear opens the module's settings, and the preset used is shown. Fix it on Manage Modules and run it again.