What Panels does
Panels is an audio data company that creates high-quality datasets for training and evaluating voice and speech AI models. It works directly with AI teams to build both off-the-shelf and custom audio datasets, handling the full pipeline of collection, quality assurance, transcription and validation in-house.
Key capabilities
Panels produces speaker-separated multilingual audio, single-speaker scripted recordings and turn-taking evaluation datasets, and designs bespoke datasets tailored to a client's needs. It iterates over time, identifying performance gaps and expanding datasets with new accents, noise conditions and edge cases to stress-test models against real-world conditions. The company also offers exclusive licensing options alongside data purchases.
Who it's for
Panels serves voice AI companies of all sizes, from early-stage startups to established research labs, that need diverse, high-quality audio data to train and evaluate speech and audio models.