Strayl
Build everywhere
21 followers
Build everywhere
21 followers
Strayl is a mobile AI coding agent - like Claude Code, but built for iOS. Edit, refactor, and preview your projects with AI. It connects to GitHub, indexes your codebase, connects to MCP servers to use their tools, and can spin up instant preview deployments.
Free Options
Launch Team / Built With
Hey Product Hunt 👋
I’m Alemzhan, founder and CEO of Strayl - a mobile AI coding agent that lets you build, refactor, and deploy right from your phone.
I started Strayl because AI coding tools like Claude Code or Cursor are powerful, but tied to the desktop. I wanted the same experience in my pocket - a full AI-assisted workspace that you can carry anywhere.
What makes it useful:
• Edit, refactor, and navigate projects with an AI coding agent.
• Connect to GitHub or create new projects directly in the app.
• Agents index your repos for deep context and smart code search.
• Connect to MCP servers and use their tools seamlessly.
• Generate temporary preview deployments for quick testing and sharing.
Strayl is starting on iOS - with Android coming next - and the goal is simple: make real AI-powered development portable.
DiffSense
Congratulations on the launch! Looks great.
Few questions I have:
- Since it's a specialized agent, I suppose the communication happens over cloud with Github? Meaning the agent changes code in a different branch or PR? And what about testing? Does the agent check for regressions or simple build checks?
- Why not instead integrate with existing coding agents like Claude or Cursor, if not already?
- I find connecting with MCP servers as something new, since there can be many long polling tasks when agent gets MCP access, this could be tracked in asyn with your mobile app.
I have heard of few platforms that offload coding operations in async mode through mobile but the MCP angle is new(atleast to me).
Kudos! Good luck.
@prastik
Hey, really appreciate the thoughtful questions!
When the agent reads a file, it caches it locally first, and all edits happen on the device. Nothing is pushed automatically. When the user taps Push, the changes are committed to the selected branch on GitHub through the API.
Right now Strayl doesn’t run linters or regression checks yet. Instead, users can instantly build and deploy projects to preview environments. The goal is to make it simple so even non-developers can see their app live right away.
We don’t connect directly with tools like Claude Code or Cursor because we plan to build our own infrastructure for AI-assisted development. Strayl runs locally and keeps all chats on the user’s device, without cloud storage. Also, we use our own indexing layer for smart contextual search across the project, and we want to keep that intelligence native to Strayl. Doing this through external providers would make it much harder to control or extend in the future. Users can still connect any model or MCP server they prefer.
For MCP tasks, everything currently runs in a simple synchronous way: the client waits for the operation to complete and then displays the result. Async tracking is on the roadmap, but for now this approach keeps things stable while we refine the agent’s workflow.
Thanks again for diving so deep into the technical side - really appreciate it!
DiffSense
@alemzhan_jakipov Thank you for the explanation. All the very best :)
Hey everyone 👋
I’m Ilyas, CTO and co-founder of Strayl.
My focus has been making Strayl feel fast, intuitive, and capable of real development on mobile - not just a chat with code.
If you’re curious, there’s a free trial so you can test how it actually works in action.
Would love your feedback 🙌
Hey Alemzhan @alemzhan_jakipov ,
Congrats on the launch of Strayl! It sounds like an awesome way to bring mobile AI-assisted development to life.
How’s the response been so far? What kind of marketing strategies are you focusing on to get Strayl into the hands of developers?