Hume AI

Hume AI provides research infrastructure for teams building and evaluating voice AI. It addresses the gap between systems that process language and generate responses and the emotional, vocal, and contextual expressions that shape how people experience conversation. The platform combines custom data collection, evaluation and simulation tooling, real-time expression measurement, and human feedback ratings, alongside public benchmarks for voice AI performance.

Its products include Data Solutions for custom data collection, the Kairos platform for building evaluation suites, simulating agent-to-agent and human-to-agent conversations, and tracking regressions, an Expression Measurement API for real-time and offline analysis across 48+ emotion categories, 50+ languages, and 600+ output metrics, and a Human Feedback API that returns per-sample scores and free-response feedback from pre-screened raters through a single API call. Public leaderboards cover speech recognition, speech understanding, text-to-speech, speech-to-speech, and automated judge performance.

The platform is intended for teams building foundation models, fine-tuning voice agents, or evaluating production voice systems. Pricing includes a free plan and paid tiers with monthly included text-to-speech characters and speech-to-speech minutes, with additional usage billed separately. Higher tiers add team seats, higher request limits, and more concurrent connections; an Enterprise edition offers custom usage, unlimited seats, and Slack support with SOC 2 Type II, GDPR, and HIPAA compliance.