Skip to main content
Coming Soon

Cognitive workspace,
built for thinkers.

Capture ideas. Discover connections.
Rediscover what you forgot.

Knowledge

Your knowledge. In one system.

Everything you've ever saved: notes, bookmarks, PDFs, email. Unified in one workspace that actually understands what you've collected.

Save anything from anywhere

Write notes with 30 block types. Clip full articles in one click. Drop a PDF and it's read and searchable instantly. Powered by a native Rust backend.

PDF
DOC
MP3
MP4
PNG
EPUB
CSV
ZIP
Chrome
YouTube
X
Reddit
Medium
Substack
NotesBookmarksPDFssheetsWeb ClipsCodeImagesVideosTasksNotesBookmarksPDFssheetsWeb ClipsCodeImagesVideosTasks

Search by meaning

Ask in plain language. Nernayam searches across notes, PDFs, and bookmarks at once, returning what you meant, not just what you typed.

"what were my notes on local AI from last month"
NoteLLM Benchmarks 202597%
PDFHugging Face paper91%
WebOllama blog post84%

Organized without effort

Smart folders update in real time. AI auto-tags every bookmark the moment it lands. No manual sorting ever.

#ai#startup#rust#research#product#design#yc#llm#reading

Separate vaults for work and life

Create multiple isolated vaults, each with its own notes, bookmarks, folders, and AI context. Nothing bleeds across. Switch in one click.

File organiser
Intelligence

Saved with meaning. Found by instinct.

Every note, PDF, and bookmark gets a semantic fingerprint the moment you save it — not keywords, but the actual idea behind the text. So connections surface on their own, and search finds what you meant.

Surfaces what you forgot

Drop a PDF. Write a note. Save a bookmark. The moment it lands, Nernayam encodes its meaning — not the words, the idea. Months later, when something related comes up, it connects automatically. No tags. No folders. No manual linking.

Attention Is All You Need.pdf
Indexed by meaning
Related in your vault
My notes on self-attention mechanisms
2 months ago
94%
Karpathy's scaling laws post
5 months ago
88%
BERT: Pre-training Deep Bidirectional Transformers.pdf
4 months ago
83%

See how your ideas connect

Every item in your vault becomes a node — linked by the ideas inside it, not just the links you drew. The graph maps both explicit connections and ones Nernayam discovered by meaning.

Knowledge graph visualization showing connected ideasConceptsResearchIdeasProjectsBookmarksGraphAI

Never lose context.

Reopen a note after six months — Nernayam shows you what you were working on, what's changed since, and every new item that now connects to it. Your vault has a memory longer than yours.

Scaling Laws for Neural LMs6 mo ago
Chinchilla scaling analysisNEW
Kaplan et al. — Scaling LawsNEW
DeepSeek-V2 Technical ReportNEW
+4 moreconnections found since last visit
Tasks

Tasks that think with you.

A full task manager that lives inside your knowledge. Create from notes, assign due dates, set priorities. Or just ask your AI.

Everything in one place

Subtasks, priorities, due dates: a complete GTD workflow that lives inside your notes, not in a separate app.

Today
2/5
Review pitch deck before Thursday
Thu2 subtasks
High
Write notes on Anthropic paper
Tue
Medium
Book flights for YC interview
Fri3 subtasks
High
Send follow-up to investors
Today1 subtask
Medium
Update landing page copy
Mon
Low

Just ask

Type it in plain language. Your AI creates tasks with due dates, priorities, and subtasks, automatically.

Remind me to review the pitch deck before Thursday
Just now
Nernayam AI
TaskReview pitch deck
DueThursday, Apr 10
Done. Task created.
Nernayam AI
Agents & Tools

AI that actually does things.

71 built-in tools across your vault and the web. Native integrations with the full Google suite, Notion, Slack, GitHub, Linear, and Jira — plus unlimited MCP servers.

Reader Agent

Searches everything at once

Hybrid search across notes, bookmarks, PDFs, live web, and YouTube transcripts.

Reader · Chat
Attention Paper
research/att…
arxiv.org
abs/1706.03…
Summarize the key ideas from these 2 sources
Thinking…
Read 2 attachments
Attention Is All You Need
Self-attention mechanisms
Researching

Pulling together notes on self-attention, comparing multi-head attention with cross-attention

Writer Agent

Organizes your whole workspace

Creates notes, manages folders and tags, and runs bulk operations in one instruction.

Writer · Chat
Research
47 notes · 3 PDFs
Scaling Laws.pdf
12 pages
Organize this folder and create a summary note
Planning 3 actions…
Created 1 note
Research Summary — Q2
Tagged 12 notes
#research#ai#scaling
Writing summary

Comparing scaling laws across Chinchilla, Kaplan et al., and DeepSeek-V2

11 Integrations + MCP

Works with your stack

Google Suite, Notion, Slack, GitHub, Linear, Jira — built in. Add any MCP server.

Gmail
Gmail
Calendar
Calendar
Drive
Drive
Docs
Docs
Sheets
Sheets
Slides
Slides
Notion
Notion
Slack
Slack
GitHub
GitHub
Linear
Linear
Jira
Jira
MCP
Model Hub

70+ AI models. Local or cloud.

Browse, install, and run LLaMA, Mistral, Gemma and more. Runs directly on your machine. Or connect Claude, GPT-4o, and Gemini with an API key. Switch anytime. No restarts.

Or browse HuggingFace directly and install any compatible model from the community without leaving the app.

LLaMA
LLaMA
Local
Mistral
Mistral
Local
Claude
Claude
Cloud
GPT-4o
GPT-4o
Cloud
Gemma
Gemma
Local
Qwen
Qwen
Local
Phi-4
Phi-4
Local
DeepSeek
DeepSeek
Local
More

…and so much more.

Local AI on your GPU or cloud AI like Claude and GPT-4o — both supported. Built with Rust for native speed. Private by default, yours to own forever.

cargo build --release
Compiling nernayam v2.0
Compiling pdf-extract v1.4
Compiling semantic-index v0.9
Finished release [optimized] 4.2s
70% smaller
Tauri v2

Built on Rust

Tauri backend — 70% smaller than Electron. PDF extraction, search, and AI inference all native.

GPU VRAM
2.1 GB6 GB
INFERENCE
42 tok/s
5× less VRAM
CUDA · Metal · CPU

GPU + CPU Inference

CUDA, Apple Silicon Metal, or CPU fallback. TurboQuant 3-bit: 5× VRAM reduction.

medium.com/article...
Clipped!
Articles · YouTube · Pages

Web Clipper

Chrome extension — save any page, article, or YouTube transcript straight to your vault.

Search notes, run commands...
K
New note
Search vault
Ask AI

OmniBox

Cmd+K universal palette — search, AI, and every command in one fuzzy-search popup.

Research24
Reading List12
This Week8
Auto-updates

Smart Folders

Saved views that auto-update in real time based on type, tag, date, or any property.

How Transformers Work

An in-depth guide to the architecture behind modern LLMs and attention...

#ai#research#llm
AI auto-tagged

AI Bookmarks

Save any URL — AI summary, auto-tags, and entity extraction. Fully local.

Heading Block
const ai = await model.run()
💡Callout with important info
+ Table · LaTeX · Mermaid · Embed...

30 Block Types

Code, tables, callouts, Mermaid diagrams, LaTeX, and rich embeds — all via slash command.

Vanilla
Midnight
Noir
Slate

4 Themes

Vanilla, Midnight, Noir, and Slate. Warm neutrals — not cold gray. Apple-grade polish.

Nernayam

You've been curating your ideas for years. Now they finally connect.

Local-first. Private by default. AI that runs on your machine.

Free forever. No credit card required.

FAQ

Frequently asked questions

Can't find what you're looking for? Reach out on Discord.


Yes. Nernayam is free for personal use with unlimited notes, vaults, and local AI. A Pro plan is available for cloud sync, extended AI models, and priority support.


Everything stays on your device by default. Notes, files, and API keys are stored locally using Tauri's secure storage (AES-256-GCM + OS keyring).

If you opt into cloud sync, data is encrypted end-to-end before leaving your machine. We never have access to your plaintext content.


Not necessarily. Nernayam ships a built-in Rust inference engine. Your models run entirely on your hardware with zero network calls.

Cloud AI providers (Anthropic, OpenAI, Google) are also available if you prefer faster responses or more capable models. You choose per conversation.


Yes. Nernayam auto-detects and migrates content from Markdown, Editor.js, BlockSuite, and TipTap formats. Just paste or import. The editor handles the conversion automatically.


Nernayam is a native desktop app built with Tauri. It runs on Windows, macOS, and Linux. It also includes a native capture overlay: Dynamic Island-style panel for clipping text and screenshots from any app, without switching windows.


Nernayam is local-first. Your notes live on your device, not our servers. API keys are encrypted with AES-256-GCM via your OS keyring. We have zero telemetry, zero tracking, and zero data collection.