Your saved posts, permanently yours.
Instagram bookmarks break when accounts vanish, rename, or go private. Archive your saved posts into a private database with permanent Cloudinary CDN sync, append-only event timelines, and in-browser vector AI.
Concrete cantilever structure in daylight. Study on brutalist overhangs and warm wood accents.
Editorial layout principles with asymmetrical multi-column baseline alignment.
Fluid keyboard navigation transitions and surface elevation hierarchy tokens.
Why native Instagram bookmarks fail
Instagram is designed for ephemeral consumption, not personal knowledge archives. When you rely on their servers, your visual references are fragile.
Instagram URLs & Bookmarks Expire
Instagram media URLs expire in days. When you revisit saved collections months later, thumbnails 404 and posts vanish permanently.
Permanent Cloudinary CDN Storage
Every saved image, carousel item, and video thumbnail is mirrored immediately to your Cloudinary storage with permanent URLs.
Creators Rename & Vanish
When an account changes handles or gets banned, bookmarks break or lead to empty profiles, making posts impossible to trace.
Append-Only Timeline & Event Audit
The tracker monitors account PKs, logs handle changes (from @old to @new), detects deletions, and alerts you on recoveries.
Zero Semantic or Visual Search
Instagram native saved search is non-existent. You are forced to manually scroll through thousands of tiles hoping to spot what you need.
Qdrant Vector AI & Face Search
Search with visual descriptions ("vintage typography poster") or find all posts containing specific faces using AI embeddings.
Locked to a Single Account Session
Managing personal, brand, and research Instagram accounts requires constant logging out and risking session invalidation.
Netflix-Style Multi-Profile Switcher
Switch between profiles in 1 click. Each profile retains its own independent session, user-agent, scraping schedules, and collections.
Engineered for permanence and intelligence
Every layer of the stack is optimized for resilient synchronization, high-speed querying, and data ownership.
Multi-Profile Architecture
Netflix-style profile switching. Manage multiple Instagram accounts in one instance, each with isolated cookies, user-agents, avatars, and private datasets.
Checkpoint-Driven Resumable Scraper
Engineered for rock-solid stability against Instagram pagination limits. Safely checkpoints progress, automatically classifies rate limits, and resumes without duplicate work.
Append-Only Event Timelines
Track every transformation over time. Detect deleted accounts, username renames, privacy flips, and badge changes automatically.
Cloudinary CDN Media Archival
Instagram media URLs expire rapidly. Automatically syncs high-resolution images, video clips, and carousels to permanent cloud CDN storage.
In-Browser & Qdrant Semantic Search
Search your visual archive with natural language. Client-side Transformer.js in-browser vector search and self-hosted Qdrant integration.
Face Detection & Facial Embeddings
Organize your archive by people. In-browser client & server face detection with TensorFlow.js and Face-API extracts facial descriptors to cluster recurring individuals.
Search by visual intent, not just keywords.
Instagram captions rarely describe the nuances inside an image. Our planned in-browser Transformer.js & self-hosted Qdrant engine generates 512-dimensional CLIP embeddings and 128-dimensional facial vectors for natural concept and person discovery.
In-Browser CLIP Vision & Text
Coming SoonClient-side local inference via Transformer.js without sending images to external AI cloud APIs.
In-Browser Face Recognition
Coming SoonFind all saved photos of a specific collaborator or creator across hundreds of saved posts using local TensorFlow.js.
Self-Hosted Qdrant Engine
Docker SupportSub-50ms cosine similarity queries on your own server for enterprise and large-scale personal archives.
From raw bookmark stream to intelligent archive
A resilient multi-stage pipeline designed for fault tolerance, data longevity, and instant search latency.
1. Ingestion Engine
Fetches saved posts via authenticated pagination. Auto-saves checkpoint tokens (checkpointMaxId) so interruptions never lose scraping state.
2. Relational / Document Store
Stores profiles, accounts, carousel media, append-only events, and username history with index-optimized query performance.
3. Media Mirroring
Asynchronously pipes high-res images and carousel slides to Cloudinary CDN, guaranteeing zero broken images or expired signatures.
4. AI Vector & Face Embeddings
Extracts CLIP multimodal embeddings & TensorFlow face vectors, indexing them into Qdrant for millisecond semantic visual queries.
5. Fast Responsive Interface
Provides instant searching, multi-criteria filtering, timeline auditing, interactive post carousels, and CSV/JSON export.
Deploy on your own server in 1 minute.
No closed SaaS subscriptions. Deploy to your Raspberry Pi, VPS, Unraid, TrueNAS, or home server with Docker Compose.
version: '3.8'
services:
app:
image: ghcr.io/gitnasr/instagram-saved-tracker:latest
ports:
- "3000:3000"
environment:
- DATABASE_URL=mongodb://mongo:27017/instagram
- QDRANT_URL=http://qdrant:6333
- CLOUDINARY_CLOUD_NAME=your_cloud_name
- CLOUDINARY_API_KEY=your_api_key
- CLOUDINARY_API_SECRET=your_api_secret
depends_on:
- mongo
restart: unless-stopped
mongo:
image: mongo:7.0
volumes:
- mongo_data:/data/db
restart: unless-stopped
volumes:
mongo_data:Everything you need to know
Have questions about data privacy, rate limiting, or vector search? We have answers.
Stop losing inspiration to expiring links and deleted posts.
Start archiving your saved Instagram posts today. Free, open source, and completely self-hosted.