AI automation + agents + weird device workflows.
I’m an AI automation and prompt enthusiast – mostly living in the space where:
- LLMs
- 0 Token Workflows
- device automation (Android/Termux, Windows, macOS)
- and tooling like OpenClaw, Token Firewall, and AutoJs
all crash into each other.
I like building systems that:
- save tokens instead of burning them
- learn once, replay forever
- can drive real devices (phones, desktops) instead of just generating text
- feel more like a personal operator than a chatbot
An LLM "token firewall" that sits between your client (OpenClaw, CLI, whatever) and your model:
- Caches answers + workflows so cache hits cost 0 tokens
- Learns reusable plans and actions
- Has a built-in chat UI
- Integrates with OpenClaw as a drop-in OpenAI-style provider
- Knows how to talk to real devices via platform "hands" (Android, desktop, etc.)
Repo: inoahaipro/token_firewall
An Android accessibility bridge that exposes a clean JSON API for controlling the phone:
- AutoJs6 script on-device + Python client in Termux
- Actions like
getScreen,getScreenText,findAndTap,tapAt,swipe,typeText,launchApp,openUrl, etc. - Designed to be wired into OpenClaw and Token Firewall, but usable standalone
Repo: inoahaipro/openclaw_bridge
- Languages / runtimes: Python, Html
- Agent / orchestration: OpenClaw, custom tools & skills
- Automation: Android + Termux, AutoJs, ADB, desktop shell workflows
- Platforms: Android, Windows, macOS
Right now I’m especially interested in:
- Turning messy prompt chains into repeatable, testable automations
- Letting agents control real devices in a reliable way
- Building infrastructure that makes AI more operational and less "just vibes"
- Roles or collaborations around:
- AI/LLM tooling and infra
- agentic systems
- automation for real devices (phones, desktops)
- Teams that like to ship, experiment, and push into weird new territory
If you’re building in this space and want someone who actually enjoys wiring AI into the real world: reach out.
You can find me here on GitHub as @inoahaipro.