Update README.md

README.md

@@ -7,4 +7,25 @@ sdk: static
 pinned: false
 ---
 
-Edit this `README.md` markdown file to author your organization card.
+# DiffuSpace
+
+**Building the next foundation model paradigm beyond autoregression.**
+
+We build diffusion language models to explore new possibilities for machine intelligence. Through parallel generation, bidirectional modeling, and iterative refinement, we aim to make language generation more flexible, efficient, and controllable.
+
+We believe intelligence should consider the whole context, navigate uncertainty, and refine possibilities into answers.
+
+## Research Directions
+
+- **Diffusion Language Models** — Advancing diffusion as a foundation for language modeling beyond token-by-token generation.
+- **Reasoning & Planning** — Exploring how iterative refinement supports complex reasoning, search, and planning.
+- **Code Generation** — Developing diffusion models for code generation and infilling.
+- **Multimodal Intelligence** — Extending diffusion language model backbones to vision-language and vision-language-action modeling.
+
+## Explore DiffuSpace
+
+🌐 [Website](https://diffuspace.ai/) · 📚 [Research](https://diffuspace.ai/research) · 💬 [Try Our Chat](https://diffuspace.ai/demos/chat/)
+
+---
+
+*Shaping structure from possibilities.*