Skip to main content
Hedra provides infrastructure for running and adapting multimodal models. Start with managed inference through the v3 API: one integration for image, video, audio, and language models, backed by a prepaid API wallet.

Make your first API request

Create a key, fund your wallet, and run a model through the v3 API.

Explore models

Choose a model, inspect its input schema, and estimate a request’s cost.

​
From inference to post-training

World Model Runtime is Hedra’s engine for multimodal workloads. Hedra Training connects model adaptation and evaluation with deployment. GPU Sandboxes provide isolated compute for custom code and pipelines. Training and GPU Sandboxes are currently research previews. See Hedra’s platform overview or contact the engineering team to discuss access and supported workloads. For bringing your own model or deploying on your own infrastructure, see deployment options.

​
Existing integrations

Building a new integration? Use v3. If you use the /web-app/public API, read Migrate from v2 to v3.