@denziideng: No internet, no cloud, no upload — you can have a private AI Agent right on your iPhone, and the results are amazing... Many AI assistants now upload data to the cloud, causing privacy and response speed issues... There's an open-source project that runs the AI Agent locally on the iPhone, no internet required, a thoughtful personal...
Summary
PhoneClaw is an open-source project that runs the AI Agent entirely locally on the iPhone, based on models like Gemma 4 and MiniCPM-V, no internet or data upload needed. It supports on-device operations such as voice, calendar, health data, ensuring privacy and fast response.
View Cached Full Text
Cached at: 06/22/26, 01:45 PM
No internet, no cloud, no upload — you can have a private AI Agent right on your iPhone, and it works amazingly well…
Many AI assistants today require uploading data to the cloud, raising privacy and response time issues…
An open-source project runs an AI Agent entirely on your iPhone, working offline. It’s a thoughtful personal assistant that supports voice input, calendar management, reminders, health data queries, and more!
It’s called PhoneClaw — an on-device iPhone AI Agent based on multiple local models including Gemma 4 and MiniCPM-V. Using Gemma 4 LiteRT-LM, it performs inference and skill execution directly on mobile devices — no cloud API or external model integration needed, and chat content never leaves your device. It helps you manage various data on your phone through natural language.
Key highlights:
- Fully local operation — data never leaves your device, ensuring privacy and security
- Natural language control of calendar, reminders, contacts, health data, etc.
- Built-in LIVE mode (real-time voice + camera visual interaction)
- Can connect to a Mac for more powerful model inference (remote inference)
- Supports Chinese and English interface and prompts
How to use: Install via TestFlight, download the models inside the app, grant necessary permissions, and start using. It also supports pairing with a Mac for more powerful inference.
For those interested in on-device iOS AI, privacy-focused computing, and mobile Agents — this project is worth checking out!
Share + bookmark, and send this to friends who care about AI Agents~
#PhoneClaw #OnDeviceAI #iOS #Agent #Gemma #OpenSource #PrivacyAI
Denzii 🕊️ (@denziideng): Found a package — it’s basically a nanny-level external tool for ordinary people to level up in the AI era! A full hardcore roadmap for free! 😱💥
Want to keep up with the AI era, but advanced math, linear algebra, and code can drive you crazy, not knowing where to start learning!
Similar Articles
@interjc: openclaw 都有手机 app 了
OpenClaw has launched native mobile apps for iOS and Android, allowing users to run AI agents on the go with features like channels, tasks, and replies.
@XAMTO_AI: Finally, someone has taken care of all the heavy lifting for agents to search the web — OpenClaw, Hermes Agent, CodeX can all use it. How absurd was it before? Twitter API charging, web scraping requiring subscriptions, Bilibili and Xiaohongshu blocking everything, making Claude Code search for info feel like raiding a dungeon...
Agent-Reach is an open-source tool that, with a single command, connects AI agents to 14 online platforms (such as Twitter, Bilibili, YouTube, etc.), solving the pain points of web scraping and API payment, allowing agents to directly obtain online information.
@FinanceYF5: As long as your laptop has 16GB of RAM, you can run a private AI locally without a subscription or API key, and even use it offline after downloading. After disabling Ollama’s cloud features, all conversations stay on your device and are not uploaded anywhere. Here’s how to do it, take note.
Introduces how to run a private AI locally on a 16GB RAM laptop using Ollama, no subscription or API key required, with offline use and privacy protection.
@xieike: do you understand what iPhone + Mac Mini M4 + Claude Opus 4.8 actually means > your autonomous agents run 24/7 at home …
A guide to setting up a local AI agent framework using iPhone, Mac Mini M4, and Claude Opus 4.8, allowing autonomous agents to run 24/7 at home, handle tasks, and improve over time.
This article systematically reviews AI Agent architecture and engineering practices, covering control flow, context engineering, tool design, memory, multi-agent organization, evaluation, tracing, and security. It is based on the OpenClaw implementation and emphasizes the critical role of Harness (testing and validation infrastructure) for system stability.
This article systematically reviews AI Agent architecture and engineering practices, covering control flow, context engineering, tool design, memory, multi-agent organization, evaluation, tracing, and security. It is based on the OpenClaw implementation and emphasizes the critical role of Harness (testing and validation infrastructure) for system stability.