Want to run your computer from anywhere? Learn how Anthropic’s new Dispatch feature works, its current success rate, and how to set up this AI agent.

We’ve gone from just 'chatting with a bot' to basically managing a whole team of specialized digital workers. It’s about being the architect of your own personal AI environment to build a system that actually works for you.
Claude Dispatch is a feature within the Claude Cowork desktop app that allows you to remote control your Mac from a mobile device. To set it up, you must install the Claude Cowork app on your Mac (the host) and the Claude app on your iPhone or iPad (the remote). Once installed, you enable "Dispatch" in the desktop settings to generate a unique connection code, which you then enter into the mobile app to create a secure one-to-one pairing.
In early research previews, the success rate for Dispatch is approximately 50%, often due to technical or environmental hurdles. For the agent to work, your Mac must be awake and reachable; if the computer goes to sleep, the connection is lost. Additionally, because the AI "sees" via screenshots, a cluttered desktop with too many open windows can overwhelm the agent. Keeping your workspace tidy and adjusting energy settings to "Prevent automatic sleeping" can significantly improve performance.
Dispatch operates within a "permissioned sandbox" to prevent unauthorized or harmful actions. For high-stakes tasks like deleting files or opening new apps, the system requires manual pre-approval. Users must proactively grant "Accessibility," "Full Disk Access," and "Screen Recording" permissions in the Mac’s System Settings while physically at the computer. Furthermore, the AI is guided by "Constitutional AI" principles, meaning it is programmed to be cautious and will often ask for clarification rather than risk making a mistake.
The Claude Code Router is an open-source proxy that acts as a traffic controller for AI requests. Instead of sending every simple query to an expensive, high-tier model like Claude Opus, the router analyzes the task and directs it to a more appropriate resource. It can route basic tasks to cheaper models like Gemini Flash or local models via Ollama, while reserving complex reasoning for top-tier models. This strategy can reduce AI API bills by 40% to 85% and provides resilience if one provider's servers go down.
The current landscape is dominated by Claude Opus, GPT-5.4, and Gemini 3.1 Pro, each with specific strengths. Claude is considered the "Specialist," excelling at production code and deep reasoning. GPT-5.4 is the "Swiss Army Knife," best suited for web automation and navigating operating systems due to its superior "Computer Use" benchmarks. Gemini 3.1 Pro is the "Value Leader," offering a massive 2-million-token context window that makes it the most efficient choice for analyzing giant documents or long files.
샌프란시스코에서 컬럼비아 대학교 동문들이 만들었습니다
"Instead of endless scrolling, I just hit play on BeFreed. It saves me so much time."
"I never knew where to start with nonfiction—BeFreed’s book lists turned into podcasts gave me a clear path."
"Perfect balance between learning and entertainment. Finished ‘Thinking, Fast and Slow’ on my commute this week."
"Crazy how much I learned while walking the dog. BeFreed = small habits → big gains."
"Reading used to feel like a chore. Now it’s just part of my lifestyle."
"Feels effortless compared to reading. I’ve finished 6 books this month already."
"BeFreed turned my guilty doomscrolling into something that feels productive and inspiring."
"BeFreed turned my commute into learning time. 20-min podcasts are perfect for finishing books I never had time for."
"BeFreed replaced my podcast queue. Imagine Spotify for books — that’s it. 🙌"
"It is great for me to learn something from the book without reading it."
"The themed book list podcasts help me connect ideas across authors—like a guided audio journey."
"Makes me feel smarter every time before going to work"
샌프란시스코에서 컬럼비아 대학교 동문들이 만들었습니다
