Naseem
<p> A native AI agent that does real work on your Mac </p> <p> <a href="https://www.producthunt.com/products/naseem-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1237551?app_id=339">Link</a> </p>
Last verified:
What is Naseem?
Naseem is an AI agent harness for macOS that moves beyond chat—it inspects and modifies your real files, runs terminal commands, debugs apps in the iOS Simulator via accessibility trees, and drives native Mac GUI apps. Built in Swift (lightweight, no Electron), it handles long coding sessions with crash-safe memory, context caching, and token-aware execution.
Naseem pricing
Pricing model: Freemium
Free (basic chat and file/terminal supervision) + Pro tier (iOS Simulator, Mac GUI automation, sub-agent delegation, Telegram remote); exact Pro price not stated
Naseem pros
- Direct file system and terminal access—no copy-pasting between chat and editor; runs real git, xcodebuild, Python, and tests
- iOS Simulator and Mac app automation via accessibility trees, not screenshots—faster, cheaper, more precise than visual guessing
- Long session durability—memory persists across app restarts, crashed runs recover with step history, context caching reduces token costs
- Multi-model vision fallback—if your chat model can't see images, Naseem hands screenshots to a vision model and describes them back
- Lightweight native architecture—written entirely in Swift, no embedded browser or Node runtime overhead
Naseem cons
- macOS only—requires a Mac; iOS development features require the iOS Simulator
- Pro tier required for key features—iOS Simulator driving, Mac GUI automation, sub-agent delegation, and Telegram remote are paid-only
- Requires your own API keys—bring your own LLM provider keys and Tavily/Brave keys for web search
Frequently asked questions about Naseem
Can I use Naseem with any AI model?
Yes. Naseem detects vision support from each provider's catalog; if your chat model can't see images, it automatically hands screenshots to a vision model and returns a full description.
What happens to my work if Naseem crashes mid-task?
Completed steps are preserved and clearly marked. Quit or crash, then continue—the agent remembers what was already done.
Do I have to copy code between the chat window and my editor?
No. Naseem works directly on your real files on disk—inspect, read, patch, and create without leaving the app.
How does it debug iOS apps?
It reads the app's accessibility tree as text and taps/types/swipes on named elements—no screenshots or guessing. Faster and far more precise.