Yes, Wrench.ai can analyze voiceover. The V.O. script is text, and it gets scored the same way any other copy or script does — against persona meta-measures or against a very specific creative. But the real power is in overlaying the V.O. analysis on top of the visual frame scores to see whether the audio and visual are reinforcing the same persona triggers or pulling in different directions.
A voiceover script is written content. Once uploaded, it gets evaluated against the same four dimensions as any creative — brand fit, audience fit, creative novelty, and contextual appropriateness — with your persona meta-measures providing the audience intelligence layer.
This tells you whether the words being spoken align with the target persona's motivations, tone preferences, and emotional triggers — independent of the visuals.
This is where it gets powerful. If you've already scored the video using the frame-slice approach, you can overlay the voiceover scoring on top to see whether the audio and visual tracks are working together or against each other.
For example: if the visuals are high-energy and action-oriented but the V.O. is measured and institutional, you'll see that misalignment reflected in the scores. The visual frames might score well for one persona dimension while the V.O. pulls toward a different one.
That audio/visual alignment analysis is a strong differentiator. Most creative testing tools evaluate audio or visual — not both simultaneously against the same persona.
Pull up the video creative in your workspace. Click the Chat with Agent button on that creative. Upload the voiceover commentary so the agent has the context of both the visual creative and the spoken content. From there, you can ask the agent specific questions about alignment, persona fit, or areas of conflict — and it scores the V.O. against the same creative it's looking at.
You're not running two separate analyses. You're giving the agent full context — visual and audio — and letting it evaluate the whole experience.