VoKoadsreelsproductsstickersexplainerstrailers

Made by your coding agent. Delivered as a file. 20+ image, video, voice and music models behind one command, priced in credits before anything runs.

$ npx skills add vokox-ai/vokox
GitHubPricing
Claude CodeCodexCursorChatGPTclaude.aiOpenClawKling 3.0Veo 3.1Seedance 2.0MiniMax H3Wan 2.5LTX-2Hailuo 2.3Seedream 4.5Nano BananaGPT Image 2Lyria 3.5MiniMax SpeechElevenLabs v3Claude CodeCodexCursorChatGPTclaude.aiOpenClawKling 3.0Veo 3.1Seedance 2.0MiniMax H3Wan 2.5LTX-2Hailuo 2.3Seedream 4.5Nano BananaGPT Image 2Lyria 3.5MiniMax SpeechElevenLabs v3

Works where you already work.

Two steps in the agent you use today. No new app, no provider keys.

Claude Code · terminal, VS Code, JetBrains, desktop app
  1. 1
    Add the skill
    $ npx skills add vokox-ai/vokox -a claude-code
  2. 2
    Connect your account
    $ claude mcp add --transport http vokox https://vokox.ai/mcp --header "Authorization: Bearer sk_…"

    Replace sk_… with an API key from your account.

Your API key and connector URL are on the account page. Treat them like a password; revoke any time.

Then just ask

“Make a 15-second vertical ad for my candle shop from ./candle.jpg. Show me the price before you run it.”

The agent writes the plan, tells you the price in credits and waits for your yes. Files land in your project or chat.

No need to invent it. The playbook is inside.

Hooks that work on TikTok, Reels and Shorts, UGC ad scripts for every niche, product shots that sell. Every guide is a ready prompt and plan. Pick one, or ask your agent to study the library and choose for you.

Scroll-stopping hooksUGC adsUnboxingProduct in handBefore / afterTalking headMemesHow-to demosStreet interviewsPodcast clipsListicles
Product shotsMacro & liquidsFood & drinksBeautyFashion lookbooksReal estateTravelFitnessApps & SaaSRestaurantsStickers & GIFsAutomotive
Browse the gallery →Make one like your favourite ReelNew guides every week. Your agent reads them through the CLI and MCP.
Members' library
  • Pattern interrupt: start mid-actionhooks
  • Three jump cuts in three secondshooks
  • POV handshooks
agent › using “Pattern interrupt: start mid-action” → plan.json: hook + 3 takes + captions, priced before it runs

Deliverables, not just clips.

Real results and the plan.json behind them. Hover a card to stop the strip and see the models behind it; click it to open its plan, then hover a step to see the moment it made.

ugc/plan.json
1{
2 "name": "coffee-ugc",
3 "budget": { "maxCredits": 150 },
4 "steps": [
5 { "id": "creator", "type": "image", "model": "seedream-4.5",
6 "prompt": "candid iPhone selfie, woman in cream knit sweater holding a matte black coffee bag, bright kitchen, window light, mid-sentence",
7 "aspectRatio": "9:16" },
8 { "id": "take", "type": "video", "model": "seedance-2-fast", "refs": ["@creator"],
9 "prompt": "quick voice note to a friend, brisk natural pace, half-smile, no big expressions, handheld sway",
10 "duration": 5, "audio": true,
11 "params": { "speech": "Okay, this is the coffee I've been telling you about. Fresh roast, every morning. Honestly, just try it." } },
12 { "id": "final", "type": "compose", "timeline": {
13 "clips": [{ "asset": "@take" }],
14 "captions": { "from": "@take", "style": "center-pop", "text": [
15 "Okay, this is the coffee",
16 "I've been telling you about.",
17 "Fresh roast every morning.",
18 "Honestly, just try it."
19 ] } } }
20 ]
21}
0.0 s5.1 s

Every model in one place.

One account, one price list in credits, no provider keys. Ask for auto:video.hq and the best value that day is picked for you.

See pricing
Video
17 models

Clips from text or a first frame, 5–15 s

Wan 2.2 Ultra Fast 480pfastWan 2.2 Ultra Fast 720pfastLTX-2 FasthqHailuo 2.3hqHailuo 2.3 FastfastMiniMax H3 Max 480pfastMiniMax H3hqSeedance 2.0 MinifastSeedance 2.0 FasthqWan 2.5hqSeedance 2.0premiumKling 3.0 StandardpremiumVeo 3.1 LitepremiumVeo 3.1 FastpremiumKling 3.0 PropremiumVeo 3.1premiumOmniHuman 1.5 (talking photo)premium
Image
8 models

Stills, edits, product shots, key frames

MiniMax image-01fastZ-Image TurbofastFLUX 2 KleinfastQwen-ImagehqSeedream 4.5hqNano Banana 2hqGPT Image 2premiumNano Banana Propremium
GIF
1 models

Seamless loops for stickers and reactions

GIF loop (Wan 2.2 480p + ffmpeg)fast
Music
3 models

Licensed background tracks

MiniMax Music 2.6hqLyria 3.5hqStable Audio 2.5hq
Voice
3 models

Voice-over in 30+ languages

KokorofastMiniMax Speech 2.8 TurbohqElevenLabs v3premium
Transcripts
1 models

Word-timed text from any audio or video

Whisper large-v3fast

Same models. Less setup, one bill, more finished.

We buy generation wholesale, run it in the cloud and hand your agent a plan format. Everything else is what the others make you do yourself.

VokoXHypitHiggsfield
Setup
one command, one minute
Node + pnpm + ffmpeg + Chrome + Python
hosted, own UI
Where it renders
in the cloud
on your laptop, 64 Chromium processes
in the cloud
Models
33 models, 7 providers
Seedance, GPT Image, Nano Banana
own catalog only
Cost before you run
estimate + budget cap
price list, no total
—
Works from
any agent, MCP, web
Claude Code, Codex
Claude, own app
Finished deliverables
ads, stickers, product sets, explainers
video clones
clips

Based on public documentation, September 2026.

What does it cost?

Real jobs, priced the way your agent prices them, next to what consumer AI platforms charge for the same generations.

What are you making?
Quality
Your agent would quote
80credits
  • 10 × Seedream 4.580
You pay
$0.64
$0.48 on an annual plan · $0.80 on the smallest pack
−14% vs market
VokoX $0.64
Market $0.75

VokoX at the Creator plan rate. Market: the median of what Higgsfield, Dreamina, Magnific, Runway, Pippit, Hypit, Krea, OpenArt, Pollo, Kling, ElevenLabs charge for the same models at 720p on their entry monthly plan, checked 26 Sep 2026.

A plan or pay as you go, all in credits.

Monthly plans from $12, top-up packs that never expire, and 100 free credits for every new account.

See pricing

Questions

How do I generate a video from Claude Code?+

Give Claude Code a tool that can call a video model, then ask for the clip in plain words: the agent writes the prompts, runs the generation and saves the file in your project. With VokoX that is one command, npx skills add vokox-ai/vokox, and then a request like “make a 10-second vertical ad from ./product.png”.

Is there an MCP server for AI image and video generation?+

Yes. Several generation services expose MCP servers, so Claude Code, Cursor, claude.ai or ChatGPT can call image and video models as tools. The VokoX MCP server at https://vokox.ai/mcp connects an agent to 20+ image, video, voice and music models with one API key and one credit balance.

How much does VokoX cost?+

You pay per generation in credits, where 1 credit is $0.01: a 5-second Seedance 2.0 Fast clip at 480p costs 50 credits and a Nano Banana 2 image costs 13. New VokoX accounts get 100 free credits; after that, credit packs start at $10 and plans at $12 a month.

Which AI video and image models can my agent use?+

An agent can use any model its tools reach, and each provider usually needs its own account and key. Through VokoX an agent gets Seedance 2.0, Kling 3.0, Veo 3.1, Hailuo 2.3, Wan 2.5, Nano Banana, Seedream 4.5 and GPT Image 2, plus MiniMax and ElevenLabs voices and MiniMax music, with no provider keys.

Can I recreate a viral TikTok or Reel with my own product?+

You can reuse the idea, not the footage: keep the hook, the shots and the pacing, and film them with your own product and people. VokoX’s Make one like this does it from a link up to 60 seconds long: it explains why the video works and rebuilds it for your product as a video or as photos.

Does it work with Codex, Cursor and ChatGPT?+

Any agent that runs shell commands can use a generation CLI, and chat apps can use an MCP connector. VokoX works in Claude Code, Codex, Cursor and OpenClaw through its skill and CLI, and in claude.ai and ChatGPT through its MCP server.

Your agent is ready. Give it a studio.

$ npx skills add vokox-ai/vokox