SuperPowers AI: "Jarvis for your phone" for non-techies, but with a heavy crypto vibe

2026-03-07 | Product Hunt | getsupers.com

A user holds a Rubik's cube, and the AI generates a custom interface with step-by-step solution instructions in real-time. The copy reads: "Custom UI generated live to meet users at their level of understanding" — this is the core selling point: the AI sees what you're doing and proactively builds an interface to help.

30-Second Quick Take

What is it?: A real-time visual AI Agent running on phones or glasses that can "see" what you're doing and proactively solve problems or automate tasks—no coding required, just use your voice.

Is it worth watching?: The concept is worth following, but use it with caution. The product hits the hottest trend of 2026: ambient AI + wearables. The team has a serial entrepreneur background (Ron Palmeri was involved in the early days of Google Voice and OpenDNS), and the underlying PhoneClaw technology can indeed automate Android phones. However, it's just launched, the iOS version is still in review, and the crypto token marketing is very aggressive. Real-world performance still needs time to be verified.

Three Key Questions

Is it for me?

• Target Audience: People who can't code but want to use AI to automate daily tasks—like letting AI manage social media, reply to messages, or organize phone tasks automatically.
• Are you the one?: If you find yourself repeating the same tasks (posting, screenshotting, jumping between apps) and don't want to learn programming, you're the target. If you're a developer, you're better off using OpenClaw or writing your own scripts for more flexibility.
• Use Cases:
  • Point your phone at an object → AI identifies it and automatically searches/orders/logs it for you.
  • Wear Meta glasses during a meeting → AI generates real-time notes and to-dos.
  • Automate social media management → An Agent posts, replies, and moderates 24/7.
  • When NOT to use it: Professional workflows requiring precise control (stick with n8n/Zapier).

Is it useful?

ROI Judgment: If you have a spare Android phone lying around, it's worth a free trial. But don't dump money into the crypto token yet—the product is too early, and the tokenomics are unclear.

Is it buzzworthy?

The "Wow" Factor:

• Real-time Visual Recognition + Auto-UI Generation: Pointing at a Rubik's cube and having the AI draw the solution steps live is a demo that definitely makes you go "wow."
• Cross-Device Support: One app that runs on phones, glasses, and even inside a Waymo (they recorded a demo inside one).

The "Aha!" Moment:

"Holy crap it works! Our app, SuperPowers AI, now automates moderating our X community every minute!" — @Viewforge (Founder Rohan Arun)

Real User Feedback:

Positive: "They're basically building Jarvis for your phone." — @spyzer, 126 likes Cautious: "Imagine fading a dev that keeps building and shipping this hard" — @dontbuytops (mostly promoting the token)

To be honest, most Twitter discussion currently centers on the $Phoneclaw token, with very few voices discussing the actual product experience. This is a signal: community growth is driven more by crypto incentives than organic word-of-mouth.

For Indie Developers

Tech Stack

• Core Engine: PhoneClaw — achieves device-level automation via Android Accessibility Service without needing root.
• Scripting Language: ClawScript (a JavaScript-based runtime scripting).
• Vision Capabilities: Vision-assisted UI targeting; analyzes screenshots to find click locations rather than relying on hard-coded coordinates.
• Device Access: Supports Meta Display Glasses / Apple Vision Pro / Android XR / SMS / FaceTime / WhatsApp video calls.
• Web Client: app.getsupers.com
• Inspiration: OpenClaw (developed by Peter Steinberger, who has since joined OpenAI).

Core Implementation

At its base is PhoneClaw, essentially an AI Agent runtime for Android. It uses the Accessibility Service to scrape screen content and gain control permissions, generating automation scripts on the fly via ClawScript. The vision component uses AI models to identify UI element positions and then executes clicks, inputs, and swipes. The scheduling system supports cron-like tasks.

In short: AI reads the screen → understands context → generates script → executes. It's similar to OpenClaw but packaged as a product for non-technical users.

Open Source Status

• Is it open source?: PhoneClaw is partially open-sourced (github.com/rohanarun/phoneclaw).
• Similar Projects: OpenClaw/Clawdbot (30K stars), Omi (open-source wearable).
• Difficulty to Replicate: Medium-High. The combo of Android Accessibility Service + AI vision isn't new, but making it stable across various apps requires handling many edge cases. The core difficulty lies in the accuracy of visual understanding. Expect 2-3 person-months for an MVP.

Business Model

• Monetization: Crypto token ($Phoneclaw) used to pay for agent usage fees.
• Pricing: Free to start; agent execution consumes tokens.
• Tokenomics: 100% of generated fees → buyback + lock; ~$300K market cap, 9.8% supply locked.
• User Base: Not disclosed; 167 PH votes.

Big Tech Risk

Very High. This space is being aggressively pursued by giants:

• Google is releasing AI glasses + Android XR in 2026.
• Meta holds 85.2% of the AI glasses market and acquired Limitless.
• Amazon acquired Bee AI.
• Apple is developing an AI wearable pin.
• OpenAI hired OpenClaw founder Peter Steinberger specifically to work on personal agents.

As a software-only solution, SuperPowers AI's moat is shallow. Once Google builds similar features natively into Android XR, the survival space for independent apps will be severely squeezed.

For Product Managers

Pain Point Analysis

• Problem Solved: Non-technical users cannot use powerful tools like Claude Code or GitHub.
• Severity: Moderate. Many people want to automate their phones but can't code, but the need for "AI watching what you do to proactively help" isn't yet widely recognized—it's a "you don't know you need it until you use it" category.

User Persona

• Core Users: Social media managers, content creators, non-techies looking to automate daily routines.
• Secondary Users: Crypto community members (attracted by token incentives).
• Scenarios: Automated social posting, cross-app workflows, visual decision support (like the Rubik's cube demo).

Feature Breakdown

Competitive Differentiation

Key Takeaways

1. The "AI sees what you do" paradigm: Proactive AI rather than reactive AI is a design shift all AI products should consider.
2. Runtime UI Generation: Dynamically generating interfaces based on user understanding is a brilliant concept applicable to many fields.
3. Video Calls (FaceTime/WhatsApp) as an Entry Point: A clever product decision that drastically lowers the barrier to entry.

For Tech Bloggers

Founder Story

• Rohan Arun (@Viewforge): BS in Computer Systems Engineering from Boston University, based in SF, serial entrepreneur.

  • Started in computer vision/3D reconstruction in 2010.
  • Founded Cheat Layer — the first company approved by OpenAI in 2021 to sell GPT-3 automations.
  • Holds US patents in 3D reconstruction and distributed computing.
  • Successfully sold the Agent Base project after about 9 months.
  • Previous projects raised a total of $3.9M.

• Ron Palmeri: Silicon Valley veteran, entrepreneur, and VC.

  • Co-founded GrandCentral → Acquired by Google → became Google Voice.
  • OpenDNS → Acquired by Cisco for $635M.
  • Founded Layer (TechCrunch Disrupt SF 2013 winner).
  • Founded MkII Ventures and Prism Skylabs.
  • Recently founded Trainspot (AI training data marketplace).

The Angle: Two serial entrepreneurs saw OpenClaw blow up (Peter Steinberger's project got 30K GitHub stars before he joined OpenAI) and decided to build a version for the rest of us. It's "OpenClaw for non-techies."

Controversies/Discussion Points

• Crypto Hype vs. Product Substance: The team mobilized the crypto community heavily for the PH launch ("Let's beat ChatGPT guys!"), and there's more talk about the $Phoneclaw token than the product itself. Is this a clever growth hack or just a crypto project in an AI wrapper?
• Privacy Boundaries of Always-on AI: Where does the data go for an AI Agent watching your screen 24/7? Using Accessibility Services means it can see everything.
• Software vs. Hardware: Not needing hardware is an advantage, but it also means no hardware lock-in—users can switch easily.

Hype Data

• PH Ranking: 167 votes, peaked at #2 (behind only ChatGPT).
• Twitter Buzz: Mostly from the team and crypto community; @spyzer's intro post got 126 likes.
• Search Trends: Too new for significant Google Trends data.

Content Suggestions

• The "Jarvis" Angle: "After the OpenClaw founder joined OpenAI, this team is building the version for everyone else."
• The Wearable War: Frame it within the 2026 AI wearable battle between Google, Meta, and Apple.

For Early Adopters

Pricing Analysis

Specific paid tiers haven't been fully detailed yet. It's unclear exactly what is locked behind the token wall.

Getting Started

• Setup Time: Claims to take "minutes."
• Learning Curve: Low (voice + visual interaction).
• Steps:
  1. Visit app.getsupers.com (Web version, works on all devices).
  2. Mac or Android users can start automating immediately.
  3. iOS/Vision Pro users need to wait for Apple's approval.
  4. If you have an old Android phone, install the PhoneClaw APK to use it as a dedicated automation device.

Pitfalls & Gripes

1. iOS Not Approved: Apple is still reviewing it, so iPhone users are stuck with the web version for now.
2. Crypto Barrier: Paying for agents with $Phoneclaw tokens is a hassle for non-crypto users.
3. Privacy Risk: Accessibility Service permission = the app sees everything on your phone.
4. Product Maturity: It's a fresh launch; expect bugs and missing features.

Security & Privacy

• Data Storage: Local vs. cloud strategy is not clearly stated.
• Permissions: Requires Android Accessibility Service (one of the highest permissions).
• Privacy Policy: Should be on the site, but detailed policies are hard to find.
• Audit: No public security audit.

Alternatives

For Investors

Market Analysis

• Market Size:
  • Ambient Intelligence: $48.29B (2026) → $126.92B (2030)
  • Ambient Computing: $44.62B (2025) → $269.68B (2033)
  • Wearable AI: $38.1B (2023) → $200.8B (2032)
• Growth: 25-27% CAGR.
• Drivers: AI glasses explosion (8.7M units in 2025, +322%), Qualcomm 3nm wearable chips, Google/Apple entry.

Competitive Landscape

Timing Analysis

• Why Now?: OpenClaw proved the demand for personal AI agents; Android XR is just starting; recent acquisitions of Limitless/Bee prove the category is validated.
• Tech Maturity: AI vision + LLM combos are usable, though reliability remains a challenge.
• Market Readiness: Consumer awareness of AI wearables is rising, though privacy acceptance for "always-on AI" is still a hurdle.

Team Background

• Rohan Arun: Serial entrepreneur, 5+ years in AI automation, multiple US patents.
• Ron Palmeri: Silicon Valley veteran, involved in Google Voice (acquired) and OpenDNS ($635M acquisition).

Funding Status

• Current Funding: No public venture records found.
• Token Path: Raised via $Phoneclaw token (~$300K market cap on Solana).
• Previous Track Record: Rohan's projects have raised a total of $3.9M.

Risk Factors

• Giants (Google/Meta/Apple) are moving aggressively into ambient AI.
• The crypto token model might hinder mainstream adoption.
• Software-only solutions lack hardware lock-in.
• Product is extremely early; PMF is not yet validated.

Conclusion

SuperPowers AI has a sexy vision—"Let the AI see what you're doing and proactively help"—but right now, it feels more like an early-stage experiment wrapped in a crypto token.

The team is strong (Ron Palmeri's resume is stellar), and the technical direction is right (ambient visual agents are the hottest track for 2026), but product maturity and growth strategy need more time. Crypto community enthusiasm does not equal Product-Market Fit.

Resource Links

2026-03-07 | Trend-Tracker v7.3