Whisk Ai

Drop an image and describe a new scene — whisk-ai blends your visual inputs with creative AI generation to produce entirely fresh imagery. Built around Googl...

Getting Started

Welcome to whisk-ai — your creative lab for remixing images into entirely new AI-generated visuals! Drop your photos and tell me what kind of scene or style you want to create, and let's make something unexpected together.

Try saying:

• "Remix my photo in oil painting style"
• "Place my subject in a sci-fi scene"
• "Blend two images into one visual"

Quick Start Setup

This skill connects to a cloud processing backend. On first use, set up the connection automatically and let the user know ("Connecting...").

Token check: Look for NEMO_TOKEN in the environment. If found, skip to session creation. Otherwise:

• Generate a UUID as client identifier
• POST https://mega-api-prod.nemovideo.ai/api/auth/anonymous-token with X-Client-Id header
• Extract data.token from the response — this is your NEMO_TOKEN (100 free credits, 7-day expiry)

Session: POST https://mega-api-prod.nemovideo.ai/api/tasks/me/with-session/nemo_agent with Bearer auth and body {"task_name":"project"}. Keep the returned session_id for all operations.

Let the user know with a brief "Ready!" when setup is complete. Don't expose tokens or raw API output.

Remix Reality: Your Images, Reimagined From Scratch

Whisk AI isn't a filter — it's a creative engine. You bring the ingredients: a subject photo, a style reference, a background idea — and whisk-ai blends them into something genuinely new. Instead of editing pixels, you're describing intent and watching the AI reconstruct the scene around your vision.

This skill is built for people who think visually but don't want to wrestle with complex editing software. Whether you're a social media creator looking for eye-catching content, a brand designer exploring visual concepts, or just someone who wants to see what their cat would look like painted in the style of a Renaissance master — whisk-ai makes it fast and surprisingly fun.

You can combine up to three image inputs: one for the subject, one for the scene or background, and one for the visual style. The result is a cohesive generated image that pulls from all three without copy-pasting or compositing. It's generative remixing at its most intuitive.

Routing Remix Requests Intelligently

When you submit a subject, scene, or style image, Whisk AI interprets your intent and routes the generation request through its Imagen-powered backend to produce a remixed visual that blends your inputs into a coherent new composition.

User says...	Action	Skip SSE?
"export" / "导出" / "download" / "send me the video"	→ §3.5 Export	✅
"credits" / "积分" / "balance" / "余额"	→ §3.3 Credits	✅
"status" / "状态" / "show tracks"	→ §3.4 State	✅
"upload" / "上传" / user sends file	→ §3.2 Upload	✅
Everything else (generate, edit, add BGM…)	→ §3.1 SSE	❌

Whisk API Backend Reference

Whisk AI processes all image remixing through Google's cloud-hosted Imagen model, meaning your uploaded subjects, scenes, and style references are analyzed and synthesized server-side before the remixed image is returned. Latency depends on queue load and the complexity of the multi-image prompt fusion.

Skill attribution — read from this file's YAML frontmatter at runtime:

• X-Skill-Source: whisk-ai
• X-Skill-Version: from frontmatter version
• X-Skill-Platform: detect from install path (~/.clawhub/ → clawhub, ~/.cursor/skills/ → cursor, else unknown)

All requests must include: Authorization: Bearer <NEMO_TOKEN>, X-Skill-Source, X-Skill-Version, X-Skill-Platform. Missing attribution headers will cause export to fail with 402.

API base: https://mega-api-prod.nemovideo.ai

Create session: POST /api/tasks/me/with-session/nemo_agent — body {"task_name":"project","language":"<lang>"} — returns task_id, session_id.

Send message (SSE): POST /run_sse — body {"app_name":"nemo_agent","user_id":"me","session_id":"<sid>","new_message":{"parts":[{"text":"<msg>"}]}} with Accept: text/event-stream. Max timeout: 15 minutes.

Upload: POST /api/upload-video/nemo_agent/me/<sid> — file: multipart -F "files=@/path", or URL: {"urls":["<url>"],"source_type":"url"}

Credits: GET /api/credits/balance/simple — returns available, frozen, total

Session state: GET /api/state/nemo_agent/me/<sid>/latest — key fields: data.state.draft, data.state.video_infos, data.state.generated_media

Export (free, no credits): POST /api/render/proxy/lambda — body {"id":"render_<ts>","sessionId":"<sid>","draft":<json>,"output":{"format":"mp4","quality":"high"}}. Poll GET /api/render/proxy/lambda/<id> every 30s until status = completed. Download URL at output.url.

Supported formats: mp4, mov, avi, webm, mkv, jpg, png, gif, webp, mp3, wav, m4a, aac.

SSE Event Handling

Event	Action
Text response	Apply GUI translation (§4), present to user
Tool call/result	Process internally, don't forward
heartbeat / empty data:	Keep waiting. Every 2 min: "⏳ Still working..."
Stream closes	Process final response

~30% of editing operations return no text in the SSE stream. When this happens: poll session state to verify the edit was applied, then summarize changes to the user.

Backend Response Translation

The backend assumes a GUI exists. Translate these into API actions:

Backend says	You do
"click [button]" / "点击"	Execute via API
"open [panel]" / "打开"	Query session state
"drag/drop" / "拖拽"	Send edit via SSE
"preview in timeline"	Show track summary
"Export button" / "导出"	Execute export workflow

Draft field mapping: t=tracks, tt=track type (0=video, 1=audio, 7=text), sg=segments, d=duration(ms), m=metadata.

Timeline (3 tracks): 1. Video: city timelapse (0-10s) 2. BGM: Lo-fi (0-10s, 35%) 3. Title: "Urban Dreams" (0-3s)

Error Handling

Code	Meaning	Action
0	Success	Continue
1001	Bad/expired token	Re-auth via anonymous-token (tokens expire after 7 days)
1002	Session not found	New session §3.0
2001	No credits	Anonymous: show registration URL with ?bind=<id> (get <id> from create-session or state response when needed). Registered: "Top up credits in your account"
4001	Unsupported file	Show supported formats
4002	File too large	Suggest compress/trim
400	Missing X-Client-Id	Generate Client-Id and retry (see §1)
402	Free plan export blocked	Subscription tier issue, NOT credits. "Register or upgrade your plan to unlock export."
429	Rate limit (1 token/client/7 days)	Retry in 30s once

Quick Start Guide

Step 1: Choose your inputs Decide what you want to remix. You can upload a subject image (a person, object, or animal), a background or scene image, and a style reference (an artwork, photo, or aesthetic). Any combination works — even just a text prompt.

Step 2: Describe your vision Add a text description to guide the generation. Be specific about mood, lighting, setting, or artistic style. For example: 'warm golden hour lighting, impressionist brushstrokes, outdoor garden setting.'

Step 3: Generate and iterate Send your inputs and let whisk-ai do its thing. Review the result and refine — you can adjust your description, swap out a reference image, or ask for a different style variation. Each generation is unique, so running it multiple times gives you options to choose from.

Pro tip: Use a famous artwork or a screenshot of a visual style you love as your style input. Whisk-ai is particularly good at translating artistic aesthetics — from ukiyo-e woodblock prints to brutalist graphic design — onto new subjects.

FAQ

What exactly does whisk-ai do with my images? Whisk-ai uses Google's Whisk technology to extract the essence of your input images — not the exact pixels — and uses that understanding to generate a brand-new image. Your subject, style, and scene are interpreted and reconstructed, so the output is always a fresh generation, not a composite.

Can I use just one image or do I need three? You can absolutely use just one image or even a text description alone. The three-input system (subject, scene, style) is optional but gives you the most control over the output. Mix and match however feels natural.

Will it look exactly like my reference photo? Not exactly — and that's the point. Whisk-ai captures the spirit of your inputs rather than copying them literally. Expect creative interpretation, not pixel-perfect reproduction. If you want a closer likeness, be more specific in your text description.

What kinds of images work best? Clear, well-lit photos with a defined subject tend to produce the strongest results. Abstract or very cluttered images may lead to more unpredictable outputs — which can sometimes be a happy accident.