Case file · № 08 · Dundal product

Waxal

On-device Wolof speech synthesis: a fine-tuned text-to-speech model with native-speaker voice cloning that runs on local hardware — no cloud round-trip in the synthesis loop. The heart of Dundal's local-language, data-sovereign AI work.

Dundal productLanguage AILabverified 2026-07-24
Case file · № 08

Waxala language that speaks for itself.

On-device Wolof speech synthesis: a fine-tuned text-to-speech model with native-speaker voice cloning that runs on local hardware — no cloud round-trip in the synthesis loop. The heart of Dundal's local-language, data-sovereign AI work.

  • On-device synthesisthe speech model loads and runs locally — Wolof audio without sending text to anyone
  • Voice cloningnative-speaker voices, cloned from recordings, speaking new text
  • Sovereignty in practicelanguage technology a Gambian institution can own outright
SynthesisWX 0088
Text"Jërëjëf" → waveform
Modelfine-tuned Wolof TTS
Runs onlocal hardware
Cloud callsnone in synthesis
On-device