Skip to main content
videos.ecohash.com

Make videos with AI. Pay by the second.

Talking-head avatars, generative video, voice cloning, transcription, word-pop subtitles, auto-reframe for TikTok/Shorts — and a full timeline editor, all in your browser. Hosted on EcoHash GPUs, billed to your wallet with an exact quote before every render.

~/render-queuequoted first
16:9Seedance 2.01080p · 12s · quoted $0.84
9:16Wan 2.2720p · 7s · $0.21
1:1Avatar · S2Vlip-synced · 15s

Voice clone

TTS · 30+ languages

V1
A1
T1

timeline · 3 tracksassistant: plan approved · applying edits

capabilities

Every step of a video, in one tab

Generation and editing live in the same project — what a model produces lands on the timeline, not in a downloads folder.

Generative video

Text-to-video, image-to-video and video-to-video on Wan 2.2, LTX and Seedance 2.0.

Up to 15 seconds per shot on the premium engine, six aspect ratios from 9:16 to 21:9, and 480p to 1080p output — pick the engine per shot, not per vendor.

Talking-head avatars

A real portrait plus a short voice clip becomes a lip-synced presenter.

Bring your own face and voice — the speech-to-video engine animates the portrait to match the audio, ready for explainers, updates and UGC-style clips.

Voice cloning & TTS

Record or upload a short reference voice, hand over a script, get narration back.

Cloned voices and stock text-to-speech share the same panel, so swapping a narrator is a dropdown change, not a re-recording session.

Captions & transcription

Speech-to-text in 30+ languages, with word-pop subtitles styled for social.

Transcripts land on the timeline as editable caption tracks — fix a word, restyle the pop animation, and auto-reframe the video for TikTok or Shorts.

Music & sound effects

Generate a score or a one-shot effect without leaving the project.

Generated audio drops straight onto its own track, where auto-duck keeps it under your dialogue.

Timeline editor + AI Assistant

A full multi-track editor in the browser, with an agent that edits it for you.

The assistant proposes a plan first, then generates media and cuts the timeline for real — every step undoable, every generation quoted before it runs.

Need a custom pipeline? ComfyUI Cloud runs your node graphs on EcoHash GPUs — renders land straight in your Studio media library.

the catalog

One account, a whole model catalog

Swap engines without swapping vendors. Every model shows what it does and what it costs — per second, per picture or per minute — before you press generate.

ByteDance Seed logo

Seedance 2.0

ByteDance Seed · Premium video engine

Text, image, video and audio conditioning — up to 15s per shot, 21:9 to 9:16, 1080p.

Alibaba Tongyi logo

Wan 2.2

Alibaba Tongyi · Video workhorse

Fast text-to-video for drafts and iterations — up to 7s, priced for volume.

Alibaba Tongyi logo

Wan 2.2 S2V

Alibaba Tongyi · Avatar / lip-sync

Drives a real portrait with a voice clip — the engine behind talking-head presenters.

Lightricks logo

LTX 2.5

Lightricks · Fast preview engine

Latency-optimized generation for drafts and previz — iterate cheaply, then commit a premium render.

hub

…and the rest

Image · speech · music · language

Image generation, voice cloning, transcription, music and tool-capable language models share the same wallet and the same per-unit pricing.

Browse the live catalog

The Model Hub is public — compare engines and prices before you even sign up. What you see there is live catalog data, not marketing. Vendor marks identify each engine’s creator and imply no endorsement.

how it works

Quote first. Render second.

The whole billing model fits in one sentence: you always know what a render costs before you commit to it.

step 1

Sign up free

Create an account and verify your email to claim your free starter credit — enough to try the engines before you spend a cent.

step 2

See the quote

Every generation shows its exact cost before it runs — on the Generate button and in the assistant's approval card. No surprise charges, ever.

step 3

Pay by the unit

Renders bill per second, images per picture, audio per minute — straight from your prepaid wallet. Stop any time; there's nothing to cancel.

Invite a friend and Refer & Earn adds credit to your wallet when they start creating.

pricing

No subscriptions. No seats.

You pay for what the models actually produce, in the unit that output is measured in — from a prepaid wallet that never auto-renews.

$/secvideo, by generated second

Every engine sets its own per-second rate, and it varies by resolution — the Model Hub card shows the floor, the Generate button shows rate × duration for the exact shot.

$/picimages, per picture

Thumbnails, stills and image-to-video source frames bill per generated picture, at the rate the image model lists.

$/minspeech, per audio minute

Voice cloning, text-to-speech and transcription bill by audio minute, so narration and captions stay predictable at any script length.

$/1Mlanguage, per million tokens

The assistant, script writing and vision models bill per million tokens — published on every language card in the Model Hub alongside the rest.

  • No subscriptions, no seats — a wallet you top up
  • The cost is on the Generate button before you run
  • Free starter credit when you verify your email
  • Refer & Earn turns an invite into wallet credit

Per-model rates are live catalog data on the public Model Hub — this page only promises the model: priced in the open, billed by the unit.

faq

Questions, answered straight

How does billing work?

You top up a prepaid wallet and pay per unit of output: video by the generated second, images per picture, speech and music per audio minute. Every generation shows its exact cost before it runs — on the Generate button and in the assistant's approval card — so the wallet never surprises you.

Is there anything free to start with?

Yes. Create an account and verify your email to claim a free starter credit. It spends like any wallet balance, so you can try generation, avatars and the editor before topping up.

Which models can I use?

Video engines like Seedance 2.0, Wan 2.2 and LTX, the Wan S2V avatar engine, plus image, voice-cloning, transcription, music and language models — all on one account. The public Model Hub lists the live catalog with each model's per-unit price.

Do I need to install anything?

No. The Studio runs entirely in your browser — generation, the multi-track timeline editor and export included. Heavy work runs on EcoHash GPUs, not your machine.

Where are my projects stored?

In your browser's local storage on the device you work from — projects don't sync across devices yet. Export finished renders (and download generated media you want to keep) to store them wherever you like.

Can I use my own face and voice?

Yes — that's the point of the avatar workflow. Upload a real portrait and a short voice clip, and the speech-to-video engine produces a lip-synced presenter. Only use faces and voices you have the rights to.

get started

From prompt to published video in one tab

Sign up, claim your starter credit, and see exactly what your first render costs before it runs.