Last updated: 2026-08-19
Roger AI first appeared on Product Hunt in March 2026 as a macOS screen-guidance app that overlays live click-by-click instructions on top of any open application using Google Gemini's multimodal models. It only guides: the user keeps their own hands on the mouse and keyboard the entire time, never handing control to the app itself.
About Roger AI
Roger AI is a screen-guidance assistant that launched on Product Hunt in March 2026. Rather than send someone away to a written tutorial or take over their mouse and keyboard like a fully autonomous agent, it overlays live instructions directly on top of whatever application is already open, similar to having a patient expert looking over their shoulder while they work. Under the hood, Roger is built by a small team operating under the Roger AI name via the TryRoger GitHub organization, where a backend repository for the planned Windows build is already public, though how much of the codebase is open source versus proprietary is not fully documented anywhere on the site. Roger is aimed at people who need to learn unfamiliar software fast: new hires ramping up on internal tools, freelancers picking up a client's stack, or anyone stuck mid-task who would otherwise go search for a tutorial. A business-facing SDK and MCP (Model Context Protocol) support are both on Roger's public roadmap, which would let guided workflows eventually connect to other agentic tools.
Pricing
Free for a 14-day trial on macOS. No paid pricing tiers have been published as of mid-2026; the company has not disclosed post-trial cost.
Key Features
- Live Screen Overlay: Roger reads the screen at roughly 1 frame per second and draws highlight boxes and tooltips over the exact element the user needs to click or type into next.
- Five-Agent Orchestration: A multi-agent system built on Google Gemini, including a dedicated task agent, tracks overall progress and confirms when the full workflow is complete, not just the current step.
- Zero-Integration Compatibility: Because Roger reasons over pixels instead of the DOM or accessibility tree, it can guide any visible application, including Excel, Figma, Photoshop, and browser tabs, with no per-app plugin.
- Interruptible, Conversational Guidance: Users can pause a walkthrough at any moment to ask a question in natural language, get a spoken or text answer, and then resume the task exactly where they stopped.
- Step Flagging for Correction: When a suggested step is wrong, the user can flag it directly in the app so the underlying guidance for that software is corrected for future sessions.
Pros
- Covers desktop creative and spreadsheet software that browser-extension-based product tour tools simply can't reach, since those only work inside web pages.
- Keeps the user's own hands on the mouse and keyboard, so the skill is actually learned instead of being outsourced to an agent that clicks on their behalf.
- Never forces a restart when a follow-up question comes up mid-task, which matters most for people who learn by asking why, not just what to click next.
Cons
- Ships only as a native macOS app as of mid-2026; Windows and Linux support are still on the roadmap, not shipped.
- Screen captures are sent to Roger's own servers for Gemini processing, and a promised bring-your-own-key privacy mode is not yet available.
- Pricing beyond the trial is completely unpublished, which makes it hard for a team to get budget approval before committing further.
Frequently Asked Questions
What are Roger AI's pricing plans in 2026?
Roger AI is free to use during a 14-day trial on macOS, and that trial is currently the only way to access it. The company has not disclosed what a subscription would cost once the trial ends, or whether one is even coming. Teams that need budget certainty should treat the trial as evaluation-only and check tryroger.xyz directly before planning a wider rollout.
What do you get on Roger AI's free tier?
Roger AI does not have a permanent free tier; instead it offers a full-featured trial on macOS with no feature restrictions during that window. There are no published seat limits or usage caps while the trial runs. What happens to access once the trial ends is not yet defined, so users should confirm current terms on tryroger.xyz before relying on it for ongoing work.
What should you use instead of Roger AI?
Scribe suits teams that want to document a workflow once and share a written how-to guide afterward, rather than get live in-the-moment guidance. Manus suits users who want an autonomous agent to execute a multi-step task on their behalf instead of being walked through doing it themselves. Loom-style screen recorders fit better when the goal is asynchronous training video rather than real-time interactive overlays.
What separates Roger AI from Scribe?
Scribe captures a workflow once and turns it into a shareable guide, which suits building a training library over time rather than live in-person help. Roger AI instead overlays instructions on the user's own screen in real time while they perform the task themselves. Scribe runs primarily in the browser, while Roger is a native macOS app that can guide any visible desktop application, not just web pages.
How do you set up Roger AI?
Getting started means downloading the native macOS app from tryroger.xyz and starting the free trial, which requires no account setup or per-app configuration. After installing, open whichever application needs walking through; Roger detects it automatically and begins overlaying guidance within moments, so the first walkthrough takes about as long as the task itself. There is no separate onboarding flow to sit through first.
Top Alternatives
- Manus: An agent finishing a multi-step task without supervision is Manus's job; learning how to do it yourself, guided live, is Roger AI's.
- Scribe: Building a shareable training library over time is what Scribe is for; live help finishing a task on someone's own screen right now is what Roger AI is for.