Your AI conversations,
supercharged.

The opportunity

• The right mouse button is the most underutilized input in computing, and it lacks innovation.
• AI models can now reliably read the structure of web pages, so accurate capture isn't a research problem anymore.
• Personal knowledge tools are becoming real infrastructure for a growing group of users. Still niche, but growing.
• Most of the AI industry is moving toward always-on screen-watching. There's room for a deliberate, local, user-controlled approach.

The problem

• Ambient AI tools watch your whole screen, which is low-signal. Deliberate capture — pointing at what matters — is more useful.
• A lot of what you want to remember is unstructured: a social post you want to come back to, a person on LinkedIn you want to reach out to later or add to a network graph, a quote from an article, a product you're considering. There's no fast way to get those into your personal system in an organized way.
• Personal AI tools don't have a great way to know what matters to you, so they have to guess.
• There's no good layer between the open web and your personal knowledge base — something that captures, understands, and organizes.

The solution

• RightClik is a Chrome extension plus a local companion app. The extension lets you right-click on things you care about; the companion app processes, organizes, and routes to destinations.
• Three gestures:
  • Single right-click opens an advanced, context-aware menu that understands the interface you're on, your behavior, and your goals.
  • Double right-click is a quick capture — no menu, saves what you pointed at using your defaults.
  • Long-press right-click opens a chat where you can ask AI about what you right-clicked, or spin up an agent to act on it.
• RightClik reads the page structure, not a screenshot. It can tell the difference between an author, a title, a quote, a tweet, a product.
• The companion app is a schema and processing layer: captures are tagged, typed (Person, Paper, Quote, Product, and so on), and organized before routing to destinations. Routing is mostly automatic, shaped by what you captured and your settings.
• Local-first. Captures live on the user's machine and only leave when routed somewhere.

What makes it different

Big Vision

• Desktop-wide right-click. Not only the browser — every file, message, and app surface could become a capture point.
• Hardware capture inputs — a parallel compartment of the platform. Alongside the right-click gesture, the same processing layer is envisioned to handle captures from AI recording devices — audio devices like Omi or Limitless for ambient and press-to-talk capture, and deliberate-capture devices like Sandbar rings. Not the initial ship, and specifics aren't fleshed out.
• A personal Schema.org. Typed entities drawn from each user's engagement with the world, queryable by their AI agents.
• The priors layer for personal AI. Agents learn what matters to a user from deliberate signals. Feeds personal AI frameworks like OpenClaw.
• Plugin ecosystem. Third-party tools register actions against the right-click — it becomes a common command surface for the personal tech stack.
• Team and role profiles. Organizations could deploy custom right-click vocabularies per role — sales, engineering, production.

What it's not

• Not a screen recorder or always-on capture tool.
• Not a general note-taking app. It's the layer between what you see and your personal knowledge management database (like Obsidian) — not a replacement for either.
• Not cloud-first. Local by architecture and by principle.
• Not primarily a chat interface. Chat is 1 mode among a few, not the product.