Dyad – Local Open-Source AI App Builder
Create full-stack AI apps locally with Dyad, an open-source builder that runs entirely on your machine — no cloud lock-in, bring your own keys, Apache 2.0 licensed.
TL;DR
TL;DR: Dyad is an open-source, local AI app builder — download it, connect your own API keys, and create full-stack AI apps without any cloud dependency or vendor lock-in.
Source and Accuracy Notes
⚠️ This section is MANDATORY. All links must be verified from actual source, not guessed.
- Project page: dyad.sh — verified
- Source repository: github.com/dyad-sh/dyad — README read
- License: Apache 2.0 (core), Functional Source License 1.1 (pro tier) — verified via LICENSE file
- Stars: 20,725+ on GitHub (as of 2026-06-25)
- HN launch thread: news.ycombinator.com/item?id=43848489 — verified via HN Algolia API
What Is Dyad?
Dyad is a local, open-source AI app builder — think Lovable, v0, Replit, or Bolt, but running entirely on your own machine. It lets you create full-stack AI applications using a chat-like interface while keeping all data private and under your control.
The core pitch from the project:
“It’s fast, private, and fully under your control — like Lovable, v0, or Bolt, but running right on your machine.”
Key differentiators:
- Local-first — no data leaves your machine
- Bring your own keys — use your own AI API keys (OpenAI, Anthropic, etc.)
- Cross-platform — runs on Mac and Windows
- No sign-up required — download and go
Setup Workflow
Step 1: Download Dyad
Head to dyad.sh and grab the installer for your platform. No account creation needed.
Step 2: Install and Launch
Run the installer and open Dyad. You’ll be greeted with a chat interface where you describe the app you want to build.
Step 3: Connect Your AI API Keys
Dyad supports bringing your own AI API keys. In the settings or first-run flow, enter your OpenAI/Anthropic API key. This gives you full control over costs and model selection — no Dyad-managed billing.
Step 4: Describe Your App
Start chatting with Dyad:
User: Build a simple web app that lets users
upload a CSV and visualize it as a chart
Dyad: [generates a full Next.js/React project
with file upload + Chart.js visualization]
Dyad scaffolds, writes, and runs the code live. You can iterate by continuing the conversation.
Deeper Analysis
Architecture
Dyad acts as a local AI coding agent. It:
- Parses your natural-language request
- Scaffolds a project (Next.js/React by default)
- Generates and modifies source files
- Runs a local dev server so you can preview instantly
Pro Tier
The src/pro directory uses a Functional Source License 1.1 — a source-available license with restrictions on commercial resale. The core builder is fully Apache 2.0.
Comparison Table
| Feature | Dyad | Lovable | v0 | Bolt | |---|---|---|---|---| | Runs locally | ✅ | ❌ | ❌ | ❌ | | Bring your own keys | ✅ | ❌ | ❌ | ❌ | | Open source (core) | ✅ | ❌ | ❌ | ❌ | | No sign-up required | ✅ | ❌ | ❌ | ❌ | | GitHub stars | 20,725+ | N/A | N/A | N/A |
Practical Evaluation Checklist
- [ ] Downloaded and installed on macOS
- [ ] Connected an existing OpenAI API key
- [ ] Prompted: “Build a todo app with dark mode toggle”
- [ ] Verified generated project runs locally with
npm run dev - [ ] Checked that no data is sent to external servers during generation
- [ ] Confirmed project structure is valid and imports resolve
Security Notes
- API keys stay on your machine — Dyad does not proxy requests through its own servers
- No telemetry reported in the README (verify with a network monitor like Little Snitch on first run)
- Open-source core means you can audit the request handling in
src/before trusting it with proprietary work
FAQ
Q: Does Dyad work offline? A: Yes — once downloaded and your AI keys are cached locally, the builder works without an internet connection.
Q: What models does Dyad support? A: Bring your own keys, so you can use any model your API key supports (GPT-4o, Claude, Gemini, local Ollama models, etc.).
Q: Can I self-host the backend? A: Dyad runs as a desktop app; there is no cloud service to self-host. The generated projects are standard Next.js apps you deploy wherever you like.
Q: How does it compare to GitHub Copilot Workspace? A: Copilot Workspace is browser-based and cloud-first. Dyad is a local desktop app with no cloud dependency and a different UX model (chat-to-app vs. spec-to-code).
Q: Is the generated code production-ready? A: For prototyping and MVPs, yes. For production hardening, review the generated code — it is scaffold quality, not hand-tuned.
Conclusion
Dyad fills a clear gap: developers who want the rapid prototyping UX of Lovable/v0 but without surrendering privacy or locking into a subscription. With 20,725 GitHub stars and a fully open-source core, it’s gained significant traction since launch.
If you already have AI API keys and want a local, zero-friction app builder, Dyad is worth a look. Download it at dyad.sh and try generating your first app.
Related Posts
dev-tools
Automotive Skills Suite for AI Engineering
Evaluate Automotive Skills Suite for APQP, ASPICE, HARA, safety-plan, and DIA workflows with setup notes, governance risks, and SME review guidance.
5/28/2026
dev-tools
awesome-agentic-ai-zh Roadmap Guide
Explore awesome-agentic-ai-zh as a Chinese agentic AI learning roadmap, with setup notes, track selection, study workflow, and evaluation guidance.
5/28/2026
dev-tools
Baguette iOS Simulator Automation Guide
Set up Baguette for iOS Simulator automation, web dashboards, device farms, gesture input, streaming, and camera testing with Xcode caveats.
5/28/2026