Privacy-First AI

AI Creator

Create Stunning Videos from Your Photos Using Just Words

Search your photo library with natural language. AI indexes every photo on-device using CLIP embeddings and face detection. Select photos with words like "sunset at the beach last summer" and compile them into beautiful videos — all without uploading a single photo to the cloud.

Learn more ↓

From photos to video in 4 steps

No editing skills needed. No cloud uploads. Just your words.

📁
Step 1

Scan Photo Library

Grant access to your photo library. AI Creator scans and catalogs every image locally on your device.

🧠
Step 2

AI Indexes (CLIP + Face)

MobileCLIP generates 512-dim embeddings. BlazeFace detects and clusters faces. All on-device.

🔍
Step 3

Search with Words

Type natural language queries like "birthday party with dad" and find matching photos instantly.

🎥
Step 4

Compile Video

Selected photos are compiled into a polished video with transitions, music, and effects using FFmpeg.

Why choose AI Creator?

Your photos stay on your device. Your memories become videos. No subscriptions, no cloud dependency.

🔒

On-Device ML

All AI processing runs locally using CoreML (iOS) and TensorFlow Lite (Android). Your photos never leave your device. Zero network calls for indexing or search.

💬

Natural Language Search

Search your photos the way you think. Type "rainy day at the cafe" or "kids playing in the park" and get semantically relevant results powered by CLIP embeddings.

👤

Face Detection

BlazeFace detects faces and clusters them automatically. Search by person — "photos with Mom" — without manual tagging. All processing stays on-device.

🏔

Scene Classification

AI automatically classifies scenes — beach, mountain, city, indoor, outdoor, food, pets. Browse your library by scene type without manual albums.

🎞

Video Templates

Choose from cinematic, slideshow, timelapse, and story templates. Each comes with transitions, music beds, and timing presets. Customizable durations from 15s to 5min.

🛡

Zero Cloud Upload

Not a single photo is uploaded to any server. Embeddings, face data, and compiled videos all live on your device. Your privacy is not a feature — it is the architecture.

Built with proven on-device technology

Flutter CoreML TFLite FFmpeg SQLite MobileCLIP

How It Works Under the Hood

graph LR A[Photo Library] -->|Scan| B[On-Device ML] B -->|MobileCLIP| C[512-dim Embeddings] B -->|BlazeFace| D[Face Detection] C --> E[SQLite Vector DB] D --> E E -->|NL Query| F[Semantic Search] F -->|Selected Photos| G[FFmpeg Compiler] G --> H[Video Output] style A fill:#0891b2,color:#fff style H fill:#0891b2,color:#fff

How AI Creator Compares

The only app that combines natural language photo search with on-device video compilation.

Feature AI Creator Google Photos Apple Memories Magisto Canva Video
On-Device Processing 100% Cloud Partial Cloud Cloud
NL Photo Search Full CLIP Basic Limited No No
Auto Video Compile Yes Basic Auto only Yes Manual
Privacy Zero upload All uploaded iCloud sync All uploaded All uploaded
Face Clustering On-device Cloud On-device No No
Pricing Free / One-time Free + storage Free (Apple only) $9.99/mo $12.99/mo
Photo Library Integration Full local Full cloud Full local Upload only Upload only

How it works

From install to your first video in under 5 minutes.

1

Install & Scan

Download the app, grant photo library access, and let AI Creator index your photos in the background. Takes 2-5 minutes for 10,000 photos.

2

Search & Select

Type what you are looking for in plain English. "Beach vacation 2024" or "photos with Sarah." Select the photos you want in your video.

3

Compile & Share

Pick a video template, add music, and compile. Your video is rendered on-device and ready to share via WhatsApp, Instagram, or save locally.

100%

On-Device

512

Dim Embeddings

30fps

Video Output

0

Data Uploaded

Coming Soon

More ways to relive your memories

☁️

Cloud Backup (Optional)

Opt-in encrypted backup of your index and compiled videos. Your choice — not ours. End-to-end encrypted.

👥

Shared Albums

Collaborate on video compilations with family and friends. Each person contributes photos from their device.

📤

Social Sharing

One-tap export to Instagram Reels, YouTube Shorts, WhatsApp Status, and more. Auto-format for each platform.

Roadmap

What's Coming Next

Building the most private and intelligent photo-to-video app.

📚
In Progress

Phase 1: Photo Indexing

MobileCLIP embeddings, BlazeFace face detection, SQLite vector storage, scene classification. Full on-device photo understanding.

🎬
Up Next

Phase 2: Video Compilation

FFmpeg-powered video rendering, transition effects, music integration, template system, 15s-5min output durations, 30fps 1080p output.

☁️
Planned

Phase 3: Cloud Sync (Optional)

End-to-end encrypted backup, cross-device sync, shared album invites. Strictly opt-in — your data stays yours by default.

🌐
Planned

Phase 4: Social Features

One-tap export to Reels, Shorts, and Status. Collaborative video projects with family. AI-suggested memory compilations.

Frequently Asked Questions

Everything you need to know about AI Creator.

Are my photos uploaded to any server?
No. AI Creator processes everything on your device. Your photos, embeddings, face data, and compiled videos never leave your phone. There are zero network calls during indexing, search, or video compilation.
Which phones are supported?
iOS 16+ (iPhone 11 and newer for optimal ML performance) and Android 10+ with at least 4GB RAM. The app uses CoreML on iOS and TensorFlow Lite on Android for on-device inference.
What video quality can I expect?
Videos are compiled at 1080p 30fps using FFmpeg. Output formats include MP4 (H.264) for maximum compatibility. Duration options range from 15 seconds (Instagram/WhatsApp) to 5 minutes (YouTube/full memories).
How much storage does it use?
The app itself is about 80MB (including ML models). The photo index uses roughly 2KB per photo (512-dim embedding + metadata). For 10,000 photos, that is about 20MB of index data. Compiled videos use standard MP4 sizes (30-100MB per minute).
How does face detection work?
BlazeFace runs on-device to detect faces in every photo. Detected faces are then clustered using embeddings, so you can search by person (e.g., "photos with Mom") without manually tagging anyone. You assign names to face clusters once, and AI Creator remembers.
Will there be a cloud backup option?
Yes, in Phase 3. Cloud backup will be strictly opt-in and end-to-end encrypted. It will allow cross-device sync and shared albums. But the app will always work fully offline — cloud is a convenience layer, not a requirement.
What photo and video formats are supported?
Input: JPEG, PNG, HEIC, HEIF, WebP, and RAW formats (DNG, CR2, ARW). The app reads whatever your phone's camera produces. Output videos are MP4 (H.264/AAC) for universal compatibility across all platforms and devices.
How much does AI Creator cost?
Early access is free. The core app (photo indexing, search, and video compilation) will remain free forever. Premium features like advanced video templates, cloud backup, and shared albums will be available as a one-time purchase — no subscriptions.

Register for Early Access

Be among the first to try AI Creator. Limited spots available.

+91

10-digit Indian mobile number

By registering, you agree to our Privacy Policy and Terms of Service.

Sidewall Financial Services Pvt Ltd

Kishore Rajendra

Founding Member

kishore@sidewall.in · kishorer747

Soumya Swain (Sam)

Founding Member

soumya@sidewall.in · soumya-swain