Dyad0.6.0
Import app and two more pro modes — Smart Context and Turbo Edits — and more!
Dyad v0.6.0 introduces experimental support for importing existing apps! It also adds two powerful new modes for Dyad Pro members that make AI edits faster and more efficient. Plus, lots of bug fixes.
Highlights
Import App (Experimental)
You can now import existing apps into Dyad. If you’ve built your app using tools like Lovable, V0, or Bolt (or anything else), you can bring them into Dyad.
Because apps vary so much, importing might not be as smooth as building from scratch — but it’s a start! Please report any bugs you encounter.
Smart Context (Pro)
Smart Context automatically includes the most relevant files from your codebase in the chat context. This makes responses more efficient and is especially useful for large projects.
Turbo Edits (Pro)
Turbo Edits offloads routine code changes to a faster, more cost-efficient model specialized in editing code.
- Up to 50% faster code generation — especially for large files.
- Up to 2× cheaper — saving your AI credits.
Not a Pro member? Become a Pro member today and get access to these new pro modes that make building apps faster and much more cost-effective than buying your own API credits.
Explicit Thinking
Dyad now plans out its steps in a visible “thinking” box. Early testing shows this improves output quality and makes the chat easier to follow at a glance.
Regressions Fixed
-
Fixed local model support – v0.5.0 accidentally broke support for local models (e.g. LM Studio and Ollama). This is now fixed — thanks to everyone who reported it on GitHub.
-
Fixed OpenAI usage for Pro members – In v0.5.0, Dyad mistakenly used a user’s OpenAI key (if available) instead of their Dyad Pro key. This is now resolved.
Other Fixes and Improvements
- Files edited or created outside of Dyad are now auto-committed when Dyad commits changes — eliminating confusing warnings.
- App folder renaming is more robust, especially on Windows.
- Added a “Keep going” suggestion to help when the AI doesn't generate a complete response.
- Fixed an error when connecting to Supabase databases with the
postgisextension enabled. - Added Claude Sonnet 3.5 and Haiku 3.5 as preset model options.
- Improved loading indicator while messages are generated.
What’s Next
We’ve updated the Roadmap — check it out to see what’s coming up!