Hedra
You have a script and a photo of your presenter, and no budget for a studio day. Hedra can turn a still image and a voice track into a talking, lip-synced avatar video, and it is also a wider platform for generating and editing video, images and audio, sold through a web studio, an API and an MCP and command line setup.
Alongside avatars and lip-sync, it covers video generation, upscaling and restoration, motion transfer, relighting and restyling, background removal, image generation and editing, speech, voice conversion, music, sound effects and audio cleanup. Language and vision models are on the same bill.
Creators, marketing teams and developers building video into their products are the main users.
How you pay
Hedra charges per output from a prepaid dollar balance, with no subscription, seats or minimum. Starting rates include video generation from 1.13 cents a second, avatar and lip-sync from 2.5 cents a second, image generation from 0.59 cents each, speech from 5 cents per 1,000 characters and music from 60 cents a minute. Language models start at $0.15 per million input tokens and $0.50 per million output tokens. Prices checked on hedra.com in October 2026.
Good to know
- These are starting rates. Each model in the catalog has its own rate and billing unit, so check the model before a large job.
- Voice cloning is free; you pay when the cloned voice generates speech.
- Reasoning tokens count as output tokens on language models.
- Dedicated serving, model training and GPU time are priced per engagement through sales.
Hedra also publishes a changelog, a status page and its research on its site.
twenty47tools does not endorse or recommend any of the tools you find in our directory.