Never lose a thought again.
100% local to your Mac.
OmniCap automatically remembers everything you copy, see, and type. Instantly search your workflow history with sub-10ms latency. Zero cloud APIs required.
Figma: Landing Page V2 Concepts
Hex code copied: #7B61FF • Dimensions: 1440x900
Stripe Checkout Integration Code
await stripe.checkout.sessions.create({ mode: 'payment' });
Weekly Sync with Engineering
"Let's make sure the MLX daemon doesn't exceed 200mb RAM."
Built for the Apple ecosystem
Designed for macOS
Built natively in Swift and Rust for uncompromising performance, battery efficiency, and zero latency.
Unlock the magic of your Mac.
Move beyond simple search. OmniCap bridges the gap between your physical environment and your digital workspace.
Background Dictation
OmniCap quietly listens and transcribes your microphone input during meetings. Never take manual notes again.
Live Draw & Vectorize
Draw directly on your screen. OmniCap instantly vectorizes your sketches and converts them into pristine digital assets.
Context Stacks
Drag and drop any file, image, or link. OmniCap automatically groups them into intelligent, searchable stacks.
Powerful features.
Zero cloud dependencies.
100% Local Processing
OmniCap runs entirely on your Mac. No data is ever sent to the cloud. Zero API keys required.
Sub-10ms Search Latency
Built in Rust and Swift, optimized for Apple Silicon to search your history instantly.
Semantic OCR Vision
Find text inside images, videos, and zoom calls. If you saw it on screen, it's searchable.
Background Dictation
Automatically transcribes your microphone input during meetings so you can search your conversations.
Native MLX Daemon
Powered by Apple's MLX framework for running massive language models with zero memory bloat.
Natural Language Queries
"Find the link John sent me last week about the Stripe integration." It just works.
The Semantic Engine.
OmniCap does not log flat text. A background pipeline transforms raw clipboard chaos into a structured, searchable neural graph.
Drop anything. Synthesize everything.
Drag a 50-page competitor PDF, an architectural screenshot, or chaotic cloud logs directly into the Command Bar. OmniCap instantly vectorizes the payload locally, grounding your Neural Chat in absolute, private context.
Synthesize
The Ambient Synthesizer runs in the background, extracting semantic meaning and summarizing context locally via Apple MLX.
Recall securely
Press ⌘+Option+V. Retrieve anything instantly via Neural Chat or fuzzy search. Everything stays heavily encrypted inside the macOS sandbox.
The Midnight Protocol
While you sleep, OmniCap autonomously structures your raw data into Wings and Rooms. Over time, the Drive Engine learns your cognitive profile, acting as a tailored co-pilot.
Context-Aware Dictation.
Stop losing your voice context. OmniCap's background daemon lets your dictation survive across application boundaries.
Start dictating an email in Superhuman.
Start dictating your code review in Slack.
The Fracture
Command-Tab to Safari to check a stat.
Dictation aggressively cuts off.
Continuous Flow
Command-Tab into Xcode to verify a function name.
The mic stays hot.
Command-Tab back to Superhuman.
Trigger shortcut again. Lose train of thought.
Keep talking.
OmniCap buffers the audio and transcribes it directly into Xcode where your cursor is. Save hours of micro-friction.
The Extensibility Engine.
OmniCap isn't just a clipboard logger. It ships with a massive registry of Python/MLX local skills that manipulate your data exactly when you need it.
Multi-File Glassmorphism
Ultra-dense, native Mac aesthetics. We handle the chaos so your screen doesn't have to. Copy 50 images or code blocks, and watch them elegantly collapse into a dense, glassmorphic UI pill.
Spatial Memory Canvas
Standard clipboards force you to scroll a decaying timeline. OmniCap maps your context spatially. Explore an infinite, pan-and-zoom node canvas where related thoughts, images, and text naturally cluster into visual networks.
Source-to-Destination Flow Tracking
The timeline stops being a list of text and becomes a Directed Graph of Thought. OmniCap tracks what you copied and exactly where you pasted it. The journey from Chrome to Xcode becomes actionable intent mapping.
People are telling us things.
Unedited feedback from early users who tried OmniCap during the beta.
I was mass-copying code blocks between Xcode and a browser doc. Closed the wrong tab. OmniCap had every single snippet. Saved me a solid hour of rewriting.
Switched from Paste after 3 years. The semantic search is on another level — I typed 'that dark gradient Sarah sent' and it actually found it. Not even close to what I had before.
The dictation feature alone is worth it. I talk through drafts while walking, come back, and everything's there. Felt weird at first having it always listening, but it's genuinely local — checked Activity Monitor, zero network calls.
We handle a lot of NDA-sensitive material. The fact that nothing leaves the machine is non-negotiable for us. OmniCap is the only tool in this category that actually delivers on that promise.
Runs the MLX daemon at like 180MB RSS on my M2 Pro. I was expecting it to be way heavier. The command bar search is genuinely sub-second even with months of history.
Used it during a full day of user interviews. Didn't take a single manual note. Searched 'participant who mentioned onboarding friction' afterwards and it pulled the exact transcript segment. Incredible.
Honestly skeptical at first — 'local AI' usually means slow and janky. This is neither. The Live Draw vectorization caught me completely off guard, didn't expect that from a clipboard manager.
I copy probably 200 things a day across Notion, Figma, Slack, and Chrome. The context stacks feature groups them automatically. It's like having a research assistant that never sleeps.
I was mass-copying code blocks between Xcode and a browser doc. Closed the wrong tab. OmniCap had every single snippet. Saved me a solid hour of rewriting.
Switched from Paste after 3 years. The semantic search is on another level — I typed 'that dark gradient Sarah sent' and it actually found it. Not even close to what I had before.
The dictation feature alone is worth it. I talk through drafts while walking, come back, and everything's there. Felt weird at first having it always listening, but it's genuinely local — checked Activity Monitor, zero network calls.
We handle a lot of NDA-sensitive material. The fact that nothing leaves the machine is non-negotiable for us. OmniCap is the only tool in this category that actually delivers on that promise.
Runs the MLX daemon at like 180MB RSS on my M2 Pro. I was expecting it to be way heavier. The command bar search is genuinely sub-second even with months of history.
Used it during a full day of user interviews. Didn't take a single manual note. Searched 'participant who mentioned onboarding friction' afterwards and it pulled the exact transcript segment. Incredible.
Honestly skeptical at first — 'local AI' usually means slow and janky. This is neither. The Live Draw vectorization caught me completely off guard, didn't expect that from a clipboard manager.
I copy probably 200 things a day across Notion, Figma, Slack, and Chrome. The context stacks feature groups them automatically. It's like having a research assistant that never sleeps.
I copy probably 200 things a day across Notion, Figma, Slack, and Chrome. The context stacks feature groups them automatically. It's like having a research assistant that never sleeps.
Honestly skeptical at first — 'local AI' usually means slow and janky. This is neither. The Live Draw vectorization caught me completely off guard, didn't expect that from a clipboard manager.
Used it during a full day of user interviews. Didn't take a single manual note. Searched 'participant who mentioned onboarding friction' afterwards and it pulled the exact transcript segment. Incredible.
Runs the MLX daemon at like 180MB RSS on my M2 Pro. I was expecting it to be way heavier. The command bar search is genuinely sub-second even with months of history.
We handle a lot of NDA-sensitive material. The fact that nothing leaves the machine is non-negotiable for us. OmniCap is the only tool in this category that actually delivers on that promise.
The dictation feature alone is worth it. I talk through drafts while walking, come back, and everything's there. Felt weird at first having it always listening, but it's genuinely local — checked Activity Monitor, zero network calls.
Switched from Paste after 3 years. The semantic search is on another level — I typed 'that dark gradient Sarah sent' and it actually found it. Not even close to what I had before.
I was mass-copying code blocks between Xcode and a browser doc. Closed the wrong tab. OmniCap had every single snippet. Saved me a solid hour of rewriting.
I copy probably 200 things a day across Notion, Figma, Slack, and Chrome. The context stacks feature groups them automatically. It's like having a research assistant that never sleeps.
Honestly skeptical at first — 'local AI' usually means slow and janky. This is neither. The Live Draw vectorization caught me completely off guard, didn't expect that from a clipboard manager.
Used it during a full day of user interviews. Didn't take a single manual note. Searched 'participant who mentioned onboarding friction' afterwards and it pulled the exact transcript segment. Incredible.
Runs the MLX daemon at like 180MB RSS on my M2 Pro. I was expecting it to be way heavier. The command bar search is genuinely sub-second even with months of history.
We handle a lot of NDA-sensitive material. The fact that nothing leaves the machine is non-negotiable for us. OmniCap is the only tool in this category that actually delivers on that promise.
The dictation feature alone is worth it. I talk through drafts while walking, come back, and everything's there. Felt weird at first having it always listening, but it's genuinely local — checked Activity Monitor, zero network calls.
Switched from Paste after 3 years. The semantic search is on another level — I typed 'that dark gradient Sarah sent' and it actually found it. Not even close to what I had before.
I was mass-copying code blocks between Xcode and a browser doc. Closed the wrong tab. OmniCap had every single snippet. Saved me a solid hour of rewriting.
Simple, one-time pricing.
Lifetime License
Buy once. Own forever.
- Unlimited local storage
- 100% private processing
- All future minor updates (v1.x)
- Use on up to 3 personal Macs
The Paranoid User Guarantee.
We built OmniCap because we were tired of cloud-based AI tools secretly uploading private clipboard history to train their models. We don't want your data. We built the architecture to ensure we couldn't take it even if we tried.
100% Airgapped Inference
Your clipboard data never touches an external server. OmniCap uses Apple's native MLX framework to run massive semantic models entirely on local silicon.
macOS Sandbox Enforced
OmniCap runs inside a strict, system-level sandbox. Even if we wanted to exfiltrate your clipboard history, the operating system physically prevents it.
Hardware Keychain
Your optional OpenRouter API keys are never stored in plain text. They are locked securely behind the Apple Secure Enclave using native Keychain APIs.
FAANG-Grade Memory Safety
Local AI usually means memory leaks. OmniCap ships with a zero-dependency native model downloader and an LRU VRAM orchestrator. Your unified memory never overflows.
Concealed Field Rejection
Unlike Microsoft Recall, OmniCap explicitly ignores secure text fields, 1Password entries, and Keychain payloads at the OS level. It cannot capture what you intend to hide.
The Inverse of
First Principles.
AI has commoditized execution. Your ability to write code or draft emails is no longer a competitive advantage. Execution is practically free. Your unique context is your only leverage.
Historically, business valued only what was documented. We prioritized workshops and finalized drafts. The chaotic, ephemeral thoughts of your daily workflow were valued at zero. They were impossible to scale.
OmniCap inverses the graph.
It systematically captures the fleeting, the unspoken, and the unorganized. It transforms your raw, subconscious workflow into your highest-leverage asset. Intelligence without telemetry. Memory without friction.
Frequently Asked Questions
Does OmniCap slow down my Mac?
Do I need to pay for an OpenAI subscription?
What types of data does it record?
How is this different from Paste or Maccy?
Command
Shift V
Zero Servers. Zero Subscriptions.
Because OmniCap runs entirely on your Mac's Neural Engine, our server costs are $0.00. We pass that directly to you. No OpenAI API keys. No monthly cloud taxes. Just absolute recall for a single lifetime price.
For Apple Silicon (M1, M2, M3, M4, M5) running macOS 14.0+