0.94vec[768]pgvector0.87geminiRAG
+
NOW WITH SEMANTIC SEARCH · POWERED BY GOOGLE GEMINI

NEVER
LOSE
KNOWLEDGE
AGAIN.

Free plan available·No credit card required·Your data stays yours
scroll
Interactive Preview

THE SPLIT-SCREEN WORKSPACE.

Click any resource card below in our interactive sandbox workspace. See the detail panel slide out, browse AI-extracted key takeaways, and test the inline Q&A chat.

memexDEMO WORKSPACE
Search...
Live Preview

All Saved Resources

3 total
articleSeeded

Vector Databases Explained: pgvector vs Pinecone vs Weaviate

Comprehensive comparison of vector DB options for production RAG systems. pgvector wins for projects already on Postgres - no infra overhead, good performance under 10M vectors.

#vector-db#rag#postgresql
youtubeSeeded

Building AI Agents from Scratch - Andrej Karpathy

Deep walkthrough of building autonomous agents using LLMs. Covers tool use, memory, planning loops, and why most agent frameworks overcomplicate the core ideas.

#ai-agents#llm#tutorial
githubSeeded

vercel/ai - AI SDK for streaming LLM responses

Official Vercel AI SDK. Handles streaming, tool calls, multi-step agents. Works with OpenAI, Anthropic, Google Gemini. 8.2k stars, TypeScript first.

#typescript#streaming#open-source

02 / INTERACTIVE INGESTION PLAYGROUND

Extract and Summarize in Real-Time.

Test our extraction engine directly from the landing page. Paste any article link or click a preset below.

Select Preset URL or Ingest Custom Link
Waiting for preset or URL submission
0

The average knowledge worker saves 47 resources per week. Remembers maybe 3.

We copy links, bookmark tabs, and file away repositories with the best intentions. But storage is useless without retrieval. Traditional hierarchies fail because search expects you to remember the exact title or date.

01 / The Ingestion

You save an article to read later. You never find it.

02 / The Search

You search your bookmarks. Nothing useful.

03 / The Exit

You open 12 tabs looking for it. Give up.

There's a better way. →
Operation Guide

How it works.

01 / INGEST

Save instantly.

Paste any URL or hit ⌘K from anywhere. We parse YouTube content, GitHub parameters, Twitter feeds, articles, or notes instantly.

https://github.com/pgvector/pgvector
GITHUB1w ago
pgvector/pgvector

Vector similarity search extension for Postgres...

NLP_ANALYSIS_PIPELINE100% SECURE
01 / Content extraction DONE
02 / AI Summary Generation DONE
03 / Compute 768 Vector EXECUTING
02 / ANALYZE

Auto-Summarized.

Our background models parse paragraphs, extract tags, and formulate clean summaries. Gemini embeds the text into a 768-dimensional mathematical map representing semantic values.

03 / RETRIEVE

Search by context.

Query the way you speak. Searching "that vector setup article" or "karpathy video" maps matching embeddings instantly - providing relevant outputs with no metadata setup.

that karpathy agent code
karpathy/llm.c github repo96% match
Advanced Capabilities

Built for how you actually think.

FEATURE 01 // SMART SEARCH

Smart Semantic Search

Retrieve by meaning, not just keywords. Powered by Gemini embeddings and pgvector, Memex finds exactly what you meant even if you can't recall the exact title or tags.

Click sample query above...94% similarity
URL_PARSING_METADATA
Input: https://pgvector.org
Click "Simulate Save" to test AI summary extraction.
FEATURE 02 // AI BRIEFINGS

Structured AI Summaries

Get instant structured summaries of saved resources. Memex breaks down documents into Executive Summaries, Key Takeaways, Action Items, and Tags automatically using Gemini Flash.

FEATURE 03 // CAPTURE

AI-Powered Knowledge Capture

Extract deep insights from any URL or media source. Jina Reader parses content automatically while Gemini extracts key entities, tags, and semantic concepts with zero manual sorting.

TAGS_EXTRACTED_BUFFER
Click "Generate Tags" to test.
pgvector vs pinecone performance
rag metadata parameters doc
FEATURE 04 // LIBRARY

Smart Knowledge Library

A unified dashboard for articles, videos, repos, and notes. Organize with cross-cutting collections, view detailed AI breakdowns, and filter by topic or tag seamlessly.

0+

Resources saved daily

0

Vector Dimensions

< 5s

Save time per resource

"Finally a tool that actually helps me remember what I've learned."

- Beta user, Software Engineer
Subscription

PRICING BUILT FOR MODERN KNOWLEDGE WORK

Start free, organize your knowledge, and unlock AI-powered search, summaries, and research workflows as your library grows.

tier 01Current

Personal

₹0 / forever

Perfect for students, researchers, and individual learners.

✓ Save up to 100 resources
✓ AI-generated summaries
✓ Automatic tagging
✓ Smart collections
✓ Basic keyword search
✓ Web article capture
✓ Personal knowledge library
✗ Semantic vector search
✗ AI resource chat
✗ Unlimited storage
Get started free
tier 02Most Popular

Pro

₹199 / month

For power users who want AI-native knowledge management.

✓ Unlimited resource saves
✓ Semantic vector search
✓ Conversational AI resource chat
✓ AI-powered research assistant
✓ Unlimited collections
✓ Priority processing
✓Export summaries & notes
✓ Advanced organization tools
✓ Future feature access
tier 03Coming Soon

Teams

₹499 / member / month

Built for engineering teams, research groups, and startups.

✓ Shared team knowledge base
✓ Collaborative collections
✓ Team semantic search
✓ Shared AI workspace
✓ Team research assistant
✓ Organization-wide insights
✓ Admin controls
✓ Activity history
✓ Priority support
All plans include: Private Knowledge Storage • End-to-End Encryption • No Ads • No Data Selling • Secure Cloud Infrastructure
Vision & Roadmap

Coming soon in Phase 2

Get Started

Your second brain starts here.

Stop losing the articles, videos, and ideas that matter. Start building a knowledge base that actually works.

Already have an account? Sign in →