Hedra

What is Hedra?

Hedra is a multimodal visual content platform designed for creators, brands, and development teams. Users can combine character images, scripts, audio, and motion instructions to create AI character videos that can speak, sing, or perform.

The platform has now evolved into an integrated workspace for images, videos, audio, and AI Agents, bringing together Hedra’s own developed models as well as various third-party models. Users can carry out tasks such as character design, voice recording, animation, and version updates within the same project.

Main functions of Hedra

• AI character video:Use a single image of a person, animal, or illustration along with audio to create lip-sync performances.
• Text-to-video:Enter a description of the scene, action, shot, and style to directly generate a short video.
• Tusheng Video:Convert static characters, product images, or concept art into dynamic shots.
• AI image generation:Create characters, products, scenes, and story visuals to provide material for subsequent videos.
• AI voice:It supports text-to-speech conversion, audio uploading, and certain voice cloning workflows.
• Reuse of characters and materials:Save frequently used characters, styles, and brand assets as Elements.
• Multiple speaker selection:In a group photo, specify which character will speak or sing along with the audio.
• AI Agent：Plan and execute the entire creative process based on a description, reference images, web pages, or brand materials.
• Multi-model workspace:Utilize images, videos, and digital human models from various suppliers on a centralized basis.
• Developer API:Create materials through a unified interface, submit generation tasks, and query results.

Role videos are synchronized with lip movements.

Hedra Avatar

Hedra Avatar focuses on speaking avatars and character performances; it allows combining an image with audio to create a video whose lip movements are synchronized with the speech. Users can also use behavioral cues to control smiling, gestures, emotions, and camera movements.

It is suitable for knowledge narration, brand endorsement roles, virtual hosts, and animated mascots. Clear images of the characters along with audio without clipping and low noise generally yield more stable results.

Hedra Omnia

Omnia combines the reference images, audio cues, and textual instructions to ensure that the facial expressions, body movements, and camera angles of the characters are more coordinated. It is particularly suitable for digital avatars or UGC advertisements that require a complete scene, dynamic backgrounds, and a sense of performance.

Speakers in a group photo

When the scene contains multiple characters, the Select Speaker feature can be used to identify the character who is supposed to speak. That character will speak or sing based on the audio provided, while the stability of the other characters depends on the selected model and the complexity of the scene.

Video, image, and audio models

The Hedra model catalog offers both proprietary models and third-party models, covering avatar creation, text-to-video generation, image-to-video generation, image creation, and voice synthesis. The list of models is updated regularly; the actual capabilities available, the input requirements, and the amount of Credits consumed are shown in real time within the workspace.

Comparison of Hedra prices and packages

Hedra offers a free trial, monthly subscriptions, additional credits, and custom enterprise pricing. The figures shown on the official pricing page represent the monthly cost in dollars; promotions, taxes, and payment methods may vary depending on the region.

The package pricing page shows a monthly billing structure; if annual payment is available for the account, the actual amount and the way in which Credits are issued shall be as indicated on the settlement page.

The official documentation states that the annual subscription plan typically provides the full amount for the entire year at once, rather than in monthly installments.

Credits rules

• New users usually receive 100 free Credits to try it out;
• Videos, images, audio, and different models share the same platform Credits;
• Videos are typically charged based on their actual duration of production, rather than being priced at a fixed rate per minute.
• Monthly subscription Credits are not carried over and are reset to the package quota for the next cycle.
• Any additional Credit Packs purchased are deducted once the subscription quota is exhausted, and can be retained across different subscription periods;
• After unsubscribing, most functions will be limited, and any remaining unused credits will need to be reactivated by subscribing again.
• The real-time consumption of each model should be checked before generation; it is not appropriate to estimate it based on old screenshots.

How to choose a package

Hedra usage tutorial

Creating talking AI characters

1. Prepare character images that are clear, positive, unobstructed, and have usage rights.
2. Record clean audio, or enter a script to select an appropriate AI voice.
3. Enter the Avatar workflow and select a model that meets the requirements in terms of duration, format, and effects.
4. Upload the character image and audio, and describe the emotions, gestures, composition, and background constraints.
5. Generate a short video to check the identity, lip movements, expressions, hands, and video stability.
6. After adjusting a single variable, regenerate it; once everything is correct, create the full version and download it.

Generate short videos from text

1. Determine the purpose of the video, its target audience, resolution, and estimated duration.
2. Describe the subject, location, actions, camera angles, lighting, style, and sound requirements.
3. The most suitable models and the amount of Credits consumed per use.
4. First, create a draft in a short time to verify the motion logic and composition.
5. Continue to generate variations for the satisfactory shots, and add reference images of characters if necessary.
6. After export, checks are performed on subtitles, brand information, facts, and copyrights.

Use an AI Agent to create brand videos

1. Provide product objectives, target audience, channels, and brand constraints.
2. Upload authorized reference images, brand kits, or product page information.
3. Have the Agent break down the steps involved in generating scripts, images, audio, and video.
4. Review the creative direction and the accuracy of the materials at each key stage.
5. Save the stable characters and styles as Elements to continue creating series versions.
6. Manually verify the product description, statements, prices, and the final specifications.

Which users are it suitable for

• Content creator:Create voiceovers, story characters, musical performances, and social videos.
• Brand and Marketing Team:Generate UGC ads, product explanations, and event videos in bulk.
• Educators:Create virtual instructors, course introductions, and multilingual knowledge videos.
• Podcast and media team:Convert audio programs into publishable video clips of characters.
• Film, television, and gaming teams:Explore characters, storyboards, performance, and concept previews.
• Developer:Integrate image, audio, and video generation functions into products via APIs.

Product advantages

• Character images, audio, and textual cues can work together to guide the performance;
• It supports a variety of characters such as real people, animals, cartoons, and brand mascots;
• Self-developed models and various third-party models are gathered in one workspace;
• The supply chain for materials, from images and audio to videos, is fairly complete;
• AI Agents can plan multi-step creation processes, reducing the need to switch between tools.
• It provides public developer documentation as well as a unified interface based on API keys;
• The enterprise version includes private deployment, SSO, and dedicated support.

Usage restrictions and precautions

• Free generation may result in watermarks and does not include commercial usage rights;
• The subscribed Credits are reset on a periodic basis, and any unused monthly quota is not carried over.
• There are many models, but the input parameters, duration, resolution, and consumption rules are not uniform.
• Long videos incur higher costs compared to short tests, and incorrect generation also typically consumes quota.
• Complex group settings, side views, obstructions, and noisy audio can reduce the stability of lip movements and identity recognition.
• The generated content may contain incorrect text, abnormal movements, and unnatural camera angles;
• Access to the remaining quota after canceling the subscription will be affected by the plan status.

Copyright, Commercial Use, and Security

The official pricing page lists commercial use as one of the benefits included in the paid packages; free downloads, on the other hand, are likely to be in a non-commercial format with watermarks. Even when subscribing to a paid package, users must ensure that they have legitimate permission to use the images, audio files, portraits, voices, and brand materials they upload.

• Obtain explicit consent before cloning someone’s voice or creating a digital avatar;
• It is forbidden to use character videos to pretend to be real people, to forge endorsements, or to commit fraud.
• The claims regarding performance, price, and effectiveness in advertisements must be true and verifiable;
• When published publicly, indicate that the content is AI-generated or synthesized in accordance with the applicable rules;
• Before uploading content featuring private individuals or commercial materials, check the team’s data policy;
• For sensitive corporate projects, private deployment and legal security review solutions can be evaluated.

API and open-source information

Hedra offers an official developer platform and API documentation; it is possible to create assets, generate images or character videos, as well as check tasks and balances using API keys. Web-based subscriptions may differ from API-based billing, and there may be differences regarding the available models and call limits – it is therefore necessary to refer to the console and interface documentation before starting development.

The Hedra core generation platform and the proprietary models it uses are not open-source projects; the official documentation repositories do not contain the source code for these models. The SDKs with similar names available on GitHub, such as those in Node and C#, are usually third-party implementations, and it is important to distinguish between the official interfaces and those maintained by the community when using them.

Frequently Asked Questions

Can Hedra be used for free?

Yes, new users usually receive 100 Credits for testing purposes. The free version may include watermarks and cannot be used for commercial purposes; for actual projects, it is recommended to choose a paid package and check the rules on the billing page.

How long of a video can Hedra generate?

The length depends on the model; typical video generation tasks produce short clips of a few seconds in duration, while some models capable of creating avatars and character animations can generate longer videos. The model selector in the workspace shows the current input constraints as well as the time limit.

Does Hedra support commercial use?

The price pages for Basic, Creator, Professional, Teams, and Enterprise all list the rights related to commercial use. Users are still responsible for handling the authorization matters regarding portraits, voices, trademarks, and third-party materials.

Will Hedra’s Credits be carried over?

Monthly subscription Credits are not carried over and are reset at the start of each cycle. Credit Packs purchased separately can be retained across cycles; the specific rules regarding their validity are indicated on the account’s billing page.

Does Hedra provide APIs?

The official documentation provides interfaces for uploading materials, creating tasks, querying projects, and checking Credits. When integrating with this system, it is necessary to use the official API Key and to keep it securely stored.