Resources // Creators & Researchers

Hume AI – Researcher Access Program (API Credits)

Offer API credits
Suits for Researchers
Updated Jun 2026
Apply ↗

The offer

  • Application-based program giving subsidized API credits to researchers using Hume's voice / expression models (amount not stated publicly)
  • Focus areas: emotion science, psychiatry/clinical (biomarkers, cognitive decline, therapy tools), and AI alignment (bias/safety); proposals welcome beyond these topics too
  • Note: Hume also runs a separate "Startup Grant Program," so don't conflate the two

Who qualifies

  • Submit the application with a research plan; aimed at researchers (academic or industry) applying Hume's models to their work
  • Credits granted on review; no public rules on amounts (the general signup free tier is $20 in credits)

Community Insights

Hume is the emotion/expression-AI specialist. Its EVI voice interface impresses developers (“very ‘Her’ vibes”) and its newer Octave TTS is faster and cheaper. The researcher program subsidizes API credits for studies in emotion science, psychiatry and AI alignment. The honest backdrop: the science of “emotion detection” is contested. Hume itself frames outputs as probabilistic measurement of how an expression is interpreted, not literal emotion-reading, and critics flag cultural bias and context-blindness. Good fit for affective-computing research; validate outputs rather than treating them as ground truth.

Best Practices (from community tips)

  • Frame your proposal around the named areas (emotion science / psychiatry / AI alignment) but apply even if your topic is adjacent. They invite it.
  • Treat outputs as expression measurement, not mind-reading. Design studies that account for the model’s probabilistic, culturally-trained nature.
  • Lean on the SDKs/Playground (Python/TS/React, low latency) to prototype before committing to a credit-heavy study.
  • Don’t confuse this with Hume’s Startup Grant Program. This track is for researchers.

Community Reviews

Positive

Negative

Neutral / Mixed