Ask your agent for a video.
Get the file.
VokoX turns your coding agent into a studio: 20+ image, video, GIF, voice and music models behind one command, assembled into finished clips. No installs, no provider accounts, no surprises on the bill.
One plan file. Everything else is handled.
Your agent writes the shot list and prompts. VokoX prices it, runs the steps in parallel on the cheapest capable model, stitches the result and returns files plus a share link.
- 1Brief
Tell your agent what you need in plain words. Attach a product photo or a face if you have one.
- 2Price
The agent drafts plan.json and asks VokoX for the exact cost in credits before spending anything. You say yes.
- 3Deliver
Images, clips, voice and music generate in parallel; compose adds music, captions and cuts. You get final.mp4.
Deliverables, not just clips.
Presets the agent already knows, with working prompts and a budget. Ask for the outcome; the shot list is its job.
Creator to camera, product b-roll, voice-over, music, captions. Vertical, ready for Reels or TikTok.
Six studio shots and two turn clips from one product photo. Same product, six sets.
Eight looping GIFs of one character, eight emotions. Telegram-ready WebM on request.
A presenter you generated once, lip-synced to your script in any language.
Every model, one price list.
Prices are fixed per model in credits; the provider behind a model can change, your bill does not. Ask for auto:video.hq and get the best value that day.
| Model | Tier | Credits | |
|---|---|---|---|
| Video | |||
| wan-2.2-fast-720pWan 2.2 Ultra Fast 720p | fast | 5 / s | |
| seedance-2-fastSeedance 2.0 Fast | hq | 10 / s | |
| kling-3-stdKling 3.0 Standard | premium | 18 / s | |
| veo-3.1-fastVeo 3.1 Fast | premium | 30 / s | |
| Image | |||
| z-imageZ-Image Turbo | fast | 2 / image | |
| seedream-4.5Seedream 4.5 | hq | 8 / image | |
| nano-banana-2Nano Banana 2 | hq | 10 / image | |
| GIF | |||
| gif-loopGIF loop (Wan 2.2 480p + ffmpeg) | fast | 20 flat | |
| Music | |||
| lyria-3.5Lyria 3.5 | hq | 15 / track | |
| Voice | |||
| minimax-ttsMiniMax Speech 2.6 Turbo | hq | 10 / 1k chars | |
Why not the others?
| VokoX | Hypit | Higgsfield | |
|---|---|---|---|
| Setup | npm i -g · 1 min | Node + pnpm + ffmpeg + Chrome + Python | hosted, own UI |
| Where it renders | cloud | your laptop (64 Chromium processes) | cloud |
| Models | 20+ across 4 providers | Seedance, GPT Image, Nano Banana | own catalog |
| Price of a 5-second Seedance clip | 50 credits | ≈ 1.4–2.4× more | ≈ 2.2× more |
| Cost before you run | estimate + budget cap | price list, no total | — |
| Failed generation | refunded automatically | — | plan-dependent |
| Works from | any agent + MCP + web | Claude Code, Codex | Claude, own app |
Competitor figures from public pricing pages, September 2026.
Credits that never expire.
Buy a pack, use it whenever. Bigger packs come with bonus credits. Every new account starts with 100 free credits.
Questions
Do I need an API key from Google or ByteDance?+
No. VokoX holds the provider accounts. You buy credits once and every model is available at a fixed credit price.
Which agents are supported?+
Anything that can run a shell command: Claude Code, Codex, Cursor, OpenClaw. claude.ai and ChatGPT connect through the MCP server. There is a web account too.
What happens when a generation fails?+
The hold is released automatically and the agent tells you. You pay only for delivered files.
Can I bring my own reference photo or face?+
Yes. Files you upload become references for image-to-video, edits and lip-sync. They stay private to your account.
Is there a free tier?+
Every new account gets 100 credits, enough for a few images and a test clip. No card needed.
How much does a video cost?+
A 5-second clip is 15–90 credits depending on the model, a finished 20-second ad with voice and music is about 240. The agent shows the exact number before it spends anything.