How does Ocno build the knowledge graph automatically?

Ocno turns the pages you open into a living map by capturing local context, titles, and relationships, then computing links on-device. The graph grows from your real browsing trail instead of a manual note-taking workflow.

What makes it different from traditional PKM tools?

Most PKM tools ask you to save, tag, rewrite, and maintain everything yourself. Ocno removes that upkeep by making structure emerge from browsing, so context and connections appear without another inbox to manage.

Is the knowledge graph private?

Yes. The visual graph and the signals used to build it stay on your device. Ocno is designed so your knowledge map does not have to become a cloud profile.

Capture -> Connect -> See

Your knowledge graph, built automatically as you browse.

Ocno turns pages, ideas, and recurring themes into a local-first knowledge graph you can revisit, inspect, and grow without manual PKM rituals.

Visual knowledge map

Local-first memory

Auto-linked context

PKM without upkeep

Chrome extension

A visual map of pages, notes, and ideas

Connections appear automatically as you browse

A second brain you can actually explore

0:00 / 0:00

Auto-linked

Related pages and ideas connect themselves

Visual graph

See structure and context instead of a flat history log

Local-first

Your knowledge map stays on your machine

No upkeep

No tagging, copying, or backlink chores required

NOW SHOWING

Automatically generated visual graph

Ocno turns pages and ideas into a living map you can inspect, traverse, and grow without manual upkeep.

Ask and inspect your graph

Automatically generated visual graph

Ocno turns pages and ideas into a living map you can inspect, traverse, and grow without manual upkeep.

Ask your brain in plain English

Query your memory naturally and jump into the right source, cluster, or connection without folders, tags, or exact keywords.

Context and recovery while you browse

Context right while you browse

Unlike traditional second-brain tools, Ocno brings relevant memory back into the page you are reading instead of pushing you into a separate workspace.

Recover what you saw before

Earlier pages, sources, and related threads resurface when they matter, so recall becomes part of browsing instead of a later search task.

Architecture diagram showing local on-device data flow

The graph is collected from your browsing trail

When you open pages, Ocno captures their local context, titles, and relationships so they can appear as nodes in your graph. You do not need to manually save, tag, or rewrite what you read to preserve it.

Connections are computed on-device

Shared entities, repeated topics, and adjacent ideas are compared locally so the graph can cluster related pages and reveal structure without sending your reading history away for processing.

Your visual knowledge graph stays private

The resulting graph lives on your device as your own knowledge layer. It is not uploaded as a cloud profile, and the structure of what you know remains on your side of the boundary.

01Representation

See structure, not just history

Turn scattered tabs into a map of concepts and sources
Understand how ideas cluster instead of recalling isolated URLs
Use visual context to recover where a thought came from
Make browsing memory explorable instead of invisible

02Connection

Let ideas link themselves

Surface related pages through meaning and proximity
See recurring themes as they reappear across sessions
Build context continuity without manual curation
Navigate sideways through adjacent knowledge, not just backward

03PKM

A second brain without upkeep

No copy-paste pipeline required to preserve useful context
No need to tag every page just to find it later
Keep a living map of your reading without another inbox to manage
Make PKM fit browsing instead of interrupting it

What enabled this?

5-10x Faster

Thanks to its compiler-optimized engine and full GPU utilization, Ocno delivers up to 5–10× higher performance compared to traditional CPU-based in-browser model execution.

Smaller models are catching up 🔥

Smaller models like Qwen3 4B (GPU and 4GB RAM needed) can achieve quality comparable to large earlier models such as GPT-4o in many tasks proven by various benchmarks.

Model Compression & Quantization

Advances in quantization and model compression drastically reduced memory requirements, making powerful models practical on everyday devices. You can now run model comparable to GPT-4o on just 3–4 GB of RAM — enabling capabilities once limited to large cloud systems.

MacBook M1 (Apple Silicon)

Designed, developed, and extensively tested on Apple Silicon devices, including MacBook Pro M1 and MacBook Air M1.
Apple Silicon unified memory architecture and high-efficiency GPU acceleration provide excellent local AI inference performance.
Newer Apple Silicon chips (M2, M3, and beyond) deliver smoother performance.
Newer chips also enable larger and more capable models.

RAM

Minimum: 8GB

Good: 16GB

Best: 32GB+

Chip

Minimum: M1

Good: M1 Pro

Best: M3 Max+

Recommended Baseline

M1 Pro or newer16GB RAMApple Silicon

MAKER

Meet the Builder

I replaced my social life with shipping.

Boris Haviar

Founder & Software Engineer

Your subscription could not be saved. Please try again.

✅ You’re on the list. We’ll email you as soon as early access opens. Want to help shape Ocno? Reply to the confirmation email with your biggest browser pain.

Capture -> Connect -> See

Your knowledge graph, built automatically as you browse.