Launched this week
Agno Builder

Agno Builder

Non-coders: Design AI agents visually. Hand off real code.

3 followers

Visit website
Can't code but have AI agent ideas? Agno Builder lets you design workflows visually, test with real AI, then export production Python code. Hand it to your dev team instead of another spec doc. From idea to code handoff—no coding required.
Agno Builder gallery image
Agno Builder gallery image
Agno Builder gallery image
Agno Builder gallery image
Agno Builder gallery image
Free
Launch tags:Productivity•Artificial Intelligence•No-Code
Launch Team / Built With
Tines
Tines
The intelligent workflow platform
Promoted

What do you think? …

Sangam Pandey
Maker
📌
Hey Product Hunt! This started as a weekend project that got out of hand. I kept watching the same thing happen: someone with a clear AI idea spends days explaining it to developers. Wireframes. Notion docs. Slack threads. They build it—it's not quite right. "No, I meant the AI should research FIRST, then summarize..." Back to the whiteboard. I thought: what if you could just... build it yourself? Not learn to code. Just visually design the AI workflow, test if it actually works, then hand your dev team the code? So I started tinkering. Nights and weekends. Now it's Agno Builder: - Drag AI agents onto a canvas - Connect them visually - Chat with them (they actually work—not mockups) - Export clean Python code when it's right It's free. It's a pet project. Nothing fancy. But if you've ever lost days to the "no I meant THIS" loop, maybe it helps. Would love to know—what AI workflow have you been wanting to build but couldn't?
mostafa kh

@sangam_pandey the "no i meant this" loop is painfully real. explaining ai workflows to devs without a visual reference always ends in multiple revisions.

being able to test the flow before handing off code is the key part. most visual builders just make diagrams, this actually runs.

whats the agent library like? can you bring your own models or is it locked to specific providers?

Sangam Pandey
Maker

@topfuelauto, great questions!

Agent Library: It's built on https://github.com/agno-agi/agno (formerly Phidata) - an open-source framework for building AI agents. Agno handles the heavy lifting: tool execution, memory/state management, context handling, and more. The visual builder is essentially a GUI that generates Agno agent configurations.

Model Support: Agno supports 23+ model providers out of the box (I want to add bring your own key)

- OpenAI (GPT-4o, GPT-4, etc.)

- Anthropic (Claude)

- Google (Gemini)

- Mistral, Cohere, Groq, Together, Ollama, and more

Currently the UI defaults to OpenAI models, but adding a provider dropdown is on the roadmap. Since it's open source, you can also extend agent_runner.py to add any provider Agno supports - it's just a few lines of code.
The goal is exactly what you said - test the flow before writing code. Build it visually, chat with it, iterate, then export when ready.