NewOur fastest Text-to-Speech model yet with 1000+ voices ・ Sub-50ms latency

Build voice AIwith people who ship

Developers around the world are building voice agents, games, and tools on Gradium. Ask questions, share what you're building, and get answers from the team training the models.

Behind the build

Builders, researchers, and creators from all over the world
building the future of voice AI together

  • @Romane

    Paris, France

    Voice game
  • @Arthur

    Berlin, Germany

    Voice agent
  • @Pao

    Rio, Brésil

    Live translation
  • @Carla

    Turin, Italie

    Audiobook
  • @Lina

    Tunis, Tunisie

    Customer support
  • @Alex

    San Francisco, USA

    Language learning

DiscordIs where it happens

Join the Discord
  • Show your work
    Show your work

    Ship something, post it in #showcase. The best projects get featured in the gallery, the newsletter, and on our socials and qualify for bonus Gradium credits.

  • Get help
    Get help

    Ask a question about streaming, turn-taking, or a weird PCM mismatch and get an answer from someone who wrote the code, researchers or engineers.

  • Get early access
    Get early access

    Beta models land in the community first. Test them, tell us what breaks, and earn API credits for feedback that helps us improve the models.

Apps builtby the community

Broskis

Photograph everyday objects to turn them into weapons in a two-player arena game. Gradium Text-to-Speech and Voice Design provide the match commentary.

Last Alibi

Investigate a murder aboard a snowbound 1931 train by speaking freely with suspects. Gradium Speech-to-Text and Text-to-Speech carry the dialogue.

AfterFlow

Ferry souls to Greek gods' shrines and speak scroll incantations to calm their anger. Gradium transcribes the words and voices the gods.

Resonance

Explore desert and ice worlds through voice-driven puzzles. Gradium transcribes spoken commands, while your vocal pitch shapes other interactions.

SonicCheck

Call out chess moves to face the Black King. Gradium streaming Speech-to-Text hears each move, and Gradium Text-to-Speech voices the King's reactions.

RAGOTS

Talk with three islanders whose relationships and rumors change with your words. Gradium transcribes what you say and gives each islander a distinct voice.

Come buildsomething loud