Do I need to install the MCP server?

No. If your agent has a terminal and Node.js 20+, the skill runs the same tools with `npx vivid-mcp` and your API key. On Claude.ai or ChatGPT, add the VIVID connector (mcp.vividai.tv) instead.

How much does it cost?

The skill is free. Generations are billed in VIVID credits at the prices shown by the live models list; the agent tells you the total and asks before spending. Editing, transcription, beat analysis and screen recording cost nothing.

Will the product look like my real product?

Save it once from a photo: VIVID locks its colours and geometry and every shot references it by name. For critical SKUs (jewellery, logos, hardware) the agent can run a visual QC against the real photo and retouch only what differs.

Can I use real people as models?

Only with their consent: a recurring model built from photos requires the person's explicit consent, recorded by VIVID. Otherwise the agent designs a fictional person from attributes (age, look, style).

Which agents does it work with?

Any SKILL.md-compatible agent: Claude Code, Codex CLI, Cursor, Gemini CLI, VS Code Copilot and more via the command line, plus Claude.ai, Claude Desktop and ChatGPT through the VIVID MCP connector.

Works with the AI Agents you already use

+22 more

+18 agents

VIVID Studio — AI product photos, video, voice & editing

Turn a product photo into a full campaign: SKU-faithful images, video, voiceover, music, captions and a beat-synced edit.

Free

Security scanned

VIVID Studio — AI product photos, video, voice & editing

Example session with this skill installed

Here's a phone photo of our new ruby ring, shot at the bench (~/Desktop/ruby-ring.jpg). I need the launch campaign: a clean studio packshot, three shots with a model wearing it, and a 9:16 reel cut to music for Instagram. Keep the ring exactly as it is. What will it cost?

Read your context and instructions
Compiled the vivid studio
Generated the image

Summary

Connected to VIVID (Pro plan, 1,679.45 credits). I priced the campaign from the live model list and planned it in six steps.

Total: 82 credits. Editing and rendering are free. I waited for your go-ahead, then ran it.

The plan I proposed

#	Step	Tool	Model / settings	Credits
1	Save the ring as a product (locks shape, stones, metal) + studio packshot	`vivid_analyze_product`	vision analysis	10
2	3 campaign shots, model wearing the ring, 4:5	`vivid_generate_image`	`nano-banana-2`, 1K, `products: ["Ruby Ring"]`	3 × 6 = 18
3	Hero clip from the best shot, 5 s, 9:16	`vivid_generate_video`	`seedance-2.0-mini`, 720p, start frame	8 × 5 = 40
4	Original soundtrack, cinematic, ~100 BPM	`vivid_generate_music`	MiniMax Music 3.0	14
5	Reel: cut to the beat, punch-in on every cut, claim at the end	`vivid_create_editor_project` + `vivid_edit_timeline`	`CUT_TO_BEAT` + `punch-in` recipe	0
6	Export MP4 1080×1920	`vivid_render_project`	browser render	0

Delivered

1. Studio packshot: from the bench photo, ring geometry and stone count unchanged.

Studio packshot

2. Campaign shots: the saved product is passed by name, so the prompt only describes the scene (green velvet, low warm key light, black strapless dress). The ring is identical in every frame.

3. Image → video: the hero still animated as a 5 s vertical clip (camera push-in and the model's head turn written as separate sentences, which keeps the stone from warping).

Image to video

4. The reel: music analysed on the server (BPM + beat grid), clips re-sliced one cut per bar with CUT_TO_BEAT, punch-in recipe applied after the cut, claim "One ring. Every light." on the last 2 s, fade-out on picture and music.

[
  {"type": "CUT_TO_BEAT", "payload": {"audioClipId": "music", "beatInterval": 4, "grid": "beats"}},
  {"type": "APPLY_EFFECT_RECIPE", "payload": {"recipe": "punch-in", "intensity": 0.6}},
  {"type": "ADD_TEXT", "payload": {"text": "One ring. Every light.", "startMs": 13000, "endMs": 15000, "x": 0.5, "y": 0.82, "color": "#E8C26A"}}
]

5. Every format: the same edit reframed to 4:5 (feed), 1:1 (shop) and 16:9 (YouTube) with SET_CANVAS_PRESET, no new credits.

Every format

Spent: 82 credits · Balance: 1,597.45 credits · All files are in your VIVID gallery with download links.

Next steps

Want a QC pass? vivid_compare_product checks each shot against your bench photo (1 credit each) and returns a fix prompt for vivid_retouch if anything drifted.
Add a voiceover + word-level captions (vivid_generate_voice → vivid_transcribe → SET_SUBTITLES, about 1 credit).
Say the word and I'll make public share links for the agency (vivid_share_asset).

1792227890428-202h.jpg

PNG · 1536×1024

Generated

Example file from a real run - the skill writes it into your workspace.

Connects securely to your tools. The creator never sees your data.

About this skill

VIVID Studio gives your agent a complete AI content studio for products and brands, powered by VIVID. Ask for a campaign in plain words; the agent plans it, tells you what it costs, and delivers the files.

What it does

Product photos — packshots, white-background listings, lifestyle scenes and ad stills from one product photo, in any aspect ratio.
The same product and person in every shot — save a product once and VIVID locks its colours and geometry; create a recurring AI model and the identity stays consistent across the whole campaign.
Fix and check — masked retouch of a single object (the rest of the image stays byte-identical), plus a visual QC that compares the render with the real SKU and returns a fix.
Product video — clips, reels and UGC-style shots from a still, on the best current models (Seedance, Kling, Veo, Sora, MiniMax, Grok and more), always picking the cheapest one that does what you asked.
Audio — voiceover in 5 TTS engines (with voice cloning), royalty-free music, transcription with word timing, BPM and beat detection.
Editing — a real multi-track timeline: cut to the beat, captions (6 animated styles), titles, transitions, speed ramps and effect recipes, 16:9 ↔ 9:16 reframing, MP4 export.
Screen recordings — scripted walkthrough videos of any website, with a smooth cursor and click ripples, ready for the editor.

Why it is different

Faithful to your real product — references are locked, and a vision QC flags any difference from the SKU.
Cost-aware by design — the agent reads live prices, announces the total and waits for your yes before spending credits.
Works everywhere — through the VIVID MCP server when connected, or straight from the terminal with npx vivid-mcp and an API key: Claude Code, Codex CLI, Cursor, Gemini CLI and any SKILL.md-compatible agent.
One skill, 34 tools — a short router plus one playbook per area (images, video, audio, editing, screen recording), so the agent only reads what the task needs.

Requirements

A VIVID account (free plan available at vividai.tv) and either the VIVID MCP connector or Node.js 20+ with a VIVID_API_KEY. Generations use VIVID credits; editing, transcription, audio analysis and screen recording are free.

How to use

Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.

Reviews

No reviews yet

Be one of the first to try it. Every listed skill passes our trust checks below.

Security scanned

Passed our 8-point scan before listing

Fresh listing

Recently published to Agensi

Free forever

No account required to browse

Trust & safety

Security scanned

Verified clean 3 days ago

Free to download with an account

Requires

Vividai

Vivid MCP

Node.js 20+

Google Chrome + ffmpeg (optional, screen recording only)

Node.js

playwright

Listed3 days ago

VIVID Studio — AI product photos, video, voice & editing