Thesys Review (2026): Pricing, Features & Honest Verdict
TLDR
Thesys is a generative UI API that lets LLMs respond with live, interactive interfaces instead of plain text. It slots into any OpenAI-compatible stack, supports React, and starts free at 100 pages/month. Best for: developers building AI copilots and agent UIs. Price: free tier, then $0.01/page. Rating: 7.6/10.
What is Thesys?
If you’ve ever built an AI-powered app, you know the pain: your LLM generates brilliant responses, but they show up as walls of text. Users want charts, forms, interactive tables, slide decks. Thesys fixes that gap with their C1 API, an OpenAI-compatible middleware layer that augments LLMs to output fully rendered, interactive UI components in realtime.
The company also created OpenUI, an open standard for generative UI that uses a compact language consuming 67% fewer tokens than equivalent JSON. That’s not just a technical flex. Fewer tokens means lower costs and faster responses, which matters when you’re building production apps at scale.
Thesys is already used by 300+ teams, and they’re a funded company with a startup program. The target audience is clear: developers building AI-native software like agents, copilots, and internal tools who want their AI to communicate through rich interfaces rather than text dumps.
Key Features
C1 API: OpenAI-Compatible Middleware
The core product is the C1 API. It sits between your app and any OpenAI-compatible LLM, intercepting responses and transforming them into interactive UI. You don’t need to rip out your existing LLM integration. Just swap the endpoint and your AI starts responding with forms, charts, tables, and more. The drop-in compatibility is genuinely impressive and removes the biggest adoption barrier.
React SDK and Custom Components
Thesys ships a React SDK that makes rendering generative UI straightforward. You get a library of built-in components (charts, tables, forms, slides), but the real power is custom component support. You can register your own React components and the LLM will use them contextually. This means the generated UI can match your app’s design system instead of looking like a generic widget.
OpenUI Standard
OpenUI is Thesys’s open specification for describing UI in a token-efficient format. It’s 67% more compact than JSON-based alternatives, which translates directly into cost savings on API calls and faster streaming responses. Whether this becomes an industry standard remains to be seen, but the technical approach is sound and the efficiency gains are real.
C1 Artifacts API
Think of this as the “Claude Artifacts” equivalent you can embed in your own product. The Artifacts API generates standalone documents like slide presentations, reports, and interactive dashboards. It’s particularly useful for building AI tools where users need polished, shareable outputs rather than conversational responses.
Tool Calls for Dynamic Data
Thesys supports tool calls that let the generated UI pull in live data. A chart isn’t just static markup. It can fetch real data, respond to user interactions, and update dynamically. This bridges the gap between “AI generated a pretty picture” and “AI built a functional interface.”
Pricing
Thesys keeps pricing simple and developer-friendly. The free tier gives you 100 pages per month, which is enough for prototyping and testing. After that, it’s $0.01 per page operation with volume discounts for heavier usage. There’s also a startup program with custom pricing for early-stage companies, and enterprise plans with SLAs and dedicated support.
The per-page pricing model is refreshingly transparent. You’re not paying for seats, features behind paywalls, or confusing credit systems. You pay for what you render. For most small to medium apps, the costs will be minimal. A product generating 10,000 interactive UI pages per month would cost around $100, which is reasonable for the value delivered.
That said, costs can scale quickly for high-volume consumer apps. If you’re building something where every user interaction triggers a page render, do the math carefully before committing.
Pros and Cons
What I Like
- OpenAI-compatible drop-in means minimal integration effort with existing LLM stacks
- OpenUI’s token efficiency (67% fewer tokens than JSON) directly reduces API costs
- Custom React component support lets you maintain design consistency across your app
- Generous free tier and transparent per-page pricing with no seat-based gotchas
- The Artifacts API opens up use cases beyond chat (reports, slides, dashboards)
What Could Be Better
- React-only SDK limits adoption for teams using Vue, Svelte, or other frameworks
- Documentation is still maturing, and some advanced patterns require digging through examples
- Per-page costs could spike unpredictably for high-volume consumer-facing apps
- OpenUI is proprietary to Thesys despite being called “open,” creating potential vendor lock-in
Alternatives
Vercel AI SDK
Vercel’s AI SDK is the closest direct competitor. It includes streaming UI support and integrates tightly with Next.js. The key difference: Vercel’s approach requires you to define UI components upfront and map them to tool calls, while Thesys generates UI dynamically from any LLM response. Vercel AI SDK is free and open-source, making it a strong choice if you’re already in the Next.js ecosystem and want full control. Thesys wins on flexibility and the ability to generate unexpected UI patterns the LLM decides are appropriate.
Streamlit
Streamlit is the go-to for Python developers who want to build data apps and AI demos quickly. It’s excellent for prototyping and internal tools, but it’s Python-only and doesn’t integrate into existing React apps the way Thesys does. If you’re a data scientist building a quick demo, Streamlit is faster to start with. If you’re a product engineer embedding generative UI into a production app, Thesys is the better fit.
Gradio
Gradio (by Hugging Face) lets you build ML model interfaces with minimal code. Like Streamlit, it’s Python-centric and focused on demos and prototypes. Gradio excels at wrapping ML models with input/output interfaces, but it’s not designed for the “LLM generates arbitrary UI in realtime” use case that Thesys targets. Gradio is best for showcasing models; Thesys is best for building products.
Verdict
Thesys is solving a real problem that every AI product team hits eventually: making LLM outputs actually interactive. The C1 API’s drop-in compatibility is smart engineering, and the token efficiency of OpenUI gives it a genuine cost advantage. The 7.6 rating reflects a product that delivers on its core promise but still has room to grow, particularly around framework support beyond React and documentation depth.
For small to medium dev teams building AI-native apps, Thesys is worth trying on the free tier. The value proposition is clearest when you’re building copilots, AI agents, or internal tools where users need to interact with AI outputs rather than just read them. If your stack is React-based and you’re tired of manually building UI for every LLM response pattern, Thesys will save you significant engineering time.
FAQ
Does Thesys work with any LLM or just OpenAI?
Thesys works with any OpenAI-compatible API. That includes OpenAI models, Anthropic (via compatible wrappers), open-source models served through tools like vLLM or Ollama, and any provider offering an OpenAI-compatible endpoint. You’re not locked into a specific model provider.
Can I use Thesys with frameworks other than React?
Currently, the SDK is React-only. If your frontend uses Vue, Svelte, Angular, or vanilla JS, you’d need to either adopt React for the generative UI portions of your app or wait for additional framework support. The team has indicated broader framework support is on the roadmap, but no timeline has been confirmed.
How does Thesys pricing compare to building generative UI yourself?
Building generative UI from scratch typically requires 2-4 weeks of engineering time for a basic implementation, plus ongoing maintenance. At $0.01 per page, Thesys pays for itself quickly unless you’re at very high volumes (millions of pages/month). The break-even calculation depends on your engineering costs versus usage volume.
Is OpenUI truly an open standard?
OpenUI is published as a specification and Thesys encourages adoption, but it was created by and is currently maintained solely by Thesys. It’s “open” in the sense that the spec is public, but there’s no independent standards body governing it. Pragmatically, this means you should evaluate it as a Thesys technology rather than a vendor-neutral standard.
| Plan | Price | Plan Features | Best For |
|---|---|---|---|
| Free | $0 | 100 pages per month | Testing and prototypes |
| Pay-as-you-go | $0.01 per page | Unlimited, volume discounts available | Production apps |
| Startup Program | Custom | Special startup-friendly pricing | Early-stage companies |
| Enterprise | Custom | SLA, dedicated support, custom terms | Large teams |
Thesys Pros & Cons
What We Like
- OpenAI-compatible drop-in means minimal integration effort with existing LLM stacks
- OpenUI token efficiency (67% fewer tokens than JSON) directly reduces API costs
- Custom React component support lets you maintain design consistency
- Generous free tier and transparent per-page pricing with no seat-based gotchas
- Artifacts API opens up use cases beyond chat like reports, slides, and dashboards
What Could Be Better
- React-only SDK limits adoption for teams using Vue, Svelte, or other frameworks
- Documentation is still maturing with some advanced patterns requiring extra digging
- Per-page costs could spike unpredictably for high-volume consumer-facing apps
- OpenUI is Thesys-controlled despite the "open" branding, creating vendor lock-in risk
Thesys FAQ
Does Thesys work with any LLM or just OpenAI?
Thesys works with any OpenAI-compatible API, including OpenAI, Anthropic (via compatible wrappers), open-source models served through vLLM or Ollama, and any provider offering an OpenAI-compatible endpoint.
Can I use Thesys with frameworks other than React?
Currently the SDK is React-only. Vue, Svelte, Angular, and vanilla JS are not supported yet. The team has indicated broader framework support is planned but no timeline has been confirmed.
How does Thesys pricing compare to building generative UI yourself?
Building generative UI from scratch typically requires 2-4 weeks of engineering time plus ongoing maintenance. At $0.01 per page, Thesys pays for itself quickly unless you are at very high volumes (millions of pages per month).
Is OpenUI truly an open standard?
OpenUI is published as a public specification but is created and maintained solely by Thesys. There is no independent standards body governing it, so it should be evaluated as a Thesys technology rather than a vendor-neutral standard.
What is Thesys?
Generative UI API that turns LLM text outputs into live, interactive interfaces in realtime. 100 free pages/month, then $0.01 per page. Used by 300+ teams building AI-native apps.
Is Thesys worth it?
Thesys solves the "LLM outputs are just text" problem with a drop-in API that generates interactive UI in realtime. Strong technical foundation with OpenUI token efficiency, but React-only SDK and evolving docs hold it back from a higher score. Best value for dev teams already building AI-native products who want rich UI without the engineering overhead. We rate it 7.6/10.






