Agent Builder by Thesys - Build AI agents that respond with UI instead of text

Hey folks, Parikshit here, co-founder of Thesys. 👋

Today, we’re launching Agent Builder 🚀

A no-code agent builder to create truly agentic AI apps that respond with interactive UI like charts, forms, cards, slides, and reports, not plain text.

No code. No workflows. No frontend work.

________

⛔️The problem:

Modern AI agents aren’t workflow scripts anymore.

They reason, adapt, and decide. Tools like Claude Co-Work show this clearly.

But most agents are still text-in, text-out.

They answer questions, then stop.

Users still interpret results, build slides, create reports, or take action.

________

❇️ The solution:

Thesys Agent Builder lets you create fully functional AI agents in minutes.

You define intent and guardrails.

The agent decides how to respond and what interface to generate.

They can return:

📊 Charts and dashboards from your data

📝 Forms users can fill and submit

🧩 Cards, tables, and interactive components

📑 Slides and reports ready to share

________

🛄 The steps:

Get started in 3 steps. No code.:

1️⃣ Connect your data

Files, URLs, databases, or tools via MCP

2️⃣ Customize your agent

Add instructions. Set tone, style, and behavior.

3️⃣ Test and Share it

Publish with a link or embed it on your website.

Your agent goes live instantly.

________

✳️ The possibilities:

What you can build :

📈 Analytics agents for visual insights

🤝 Website copilots for guidance and support

💰 Sales agents that improve conversions

🛒 Ecommerce copilots for discovery and comparison

All without writing code or managing UI.

________

Under the hood, Agent Builder is powered by C1, our Generative UI engine that handles UI generation, orchestration, and runtime, so AI apps actually help users get work done.

🎁 Launch bonus: Free credits to build your first agent

👉 https://www.thesys.dev/agent-builder

Would love your feedback.

Excited to see what you build 🙌

This is what Claude Artifact Builder should have been but apparently they abandoned the project. Glad to see someone pushing this!

The "respond with UI instead of text" approach is exactly what conversational AI needs. I'm curious about how UI state is managed during multi-turn conversations - when a user modifies a previously generated chart or form, does C1 perform incremental updates on the existing component or regenerate from scratch? Also, is there support for preserving partial UI state when the conversation context shifts?

so useful

AI agents responding with actual UI instead of text walls?? this is the future fr

Interactive UI is a big leap. How do you ensure consistency so the same intent doesn’t produce wildly different interfaces across runs?

Looks promising!!

I think this wull be the future of UI! Traditional UIs will transition to chat based tools with enhanced UI components.

Cool concept!