Eggy Party launches AI-native UGC editor system
💡See how a top-tier mobile game is using LLMs and 3D generation to turn 50 million users into game developers.
⚡ 30-Second TL;DR
What Changed
Launched 'Gemini' editor system covering both frame and state synchronization.
Why It Matters
This marks a shift in UGC platforms from simple toolsets to AI-native development environments, potentially democratizing game development for millions of users.
What To Do Next
Explore how to integrate LLM-based code assistants into your own low-code/no-code platforms to improve user retention and content complexity.
Key Points
- •Launched 'Gemini' editor system covering both frame and state synchronization.
- •Integrated DeepSeek V4 Pro, GLM-5.2, and Kimi K2.7 for coding and logic assistance.
- •Utilizes Rodin 3D generation for text-to-asset creation.
- •Supports transition from visual 'Eggy Code' to Lua scripting via AI guidance.
🧠 Deep Insight
AI-generated analysis for this event.
🔑 Enhanced Key Takeaways
- •The Gemini editor system incorporates a proprietary 'Eggy-World' knowledge graph that fine-tunes the integrated LLMs specifically on NetEase's internal game engine documentation and API structures.
- •NetEase has implemented a tiered 'AI-Compute Credit' system, allowing creators to earn tokens for high-quality asset generation by participating in community moderation and tutorial creation.
- •The integration includes a real-time 'AI-Safety Guardrail' layer that automatically scans generated Lua scripts for malicious code or unauthorized API calls before deployment to the public server.
- •Data indicates that the 'Gemini' system has reduced the average time-to-publish for first-time creators by approximately 65% during the initial beta testing phase.
- •The platform now features a 'Collaborative AI' mode where multiple users can prompt the same LLM instance simultaneously to build complex multiplayer game logic in a shared workspace.
📊 Competitor Analysis▸ Show
| Feature | Eggy Party (Gemini) | Roblox (Assistant) | Rec Room (AI Studio) |
|---|---|---|---|
| Core LLM | DeepSeek/Kimi/GLM | Proprietary/Custom | Third-party API |
| 3D Generation | Rodin (Native) | Limited/External | External Plugins |
| Scripting | Visual to Lua | Luau | Custom Visual |
| Pricing | Free (Credit-based) | Subscription/Robux | Freemium |
🛠️ Technical Deep Dive
- The system utilizes a multi-agent architecture where the 'Architect Agent' handles scene layout, the 'Logic Agent' manages Lua script generation, and the 'Asset Agent' interfaces with the Rodin 3D pipeline.
- Frame and state synchronization is achieved through a custom NetEase-developed protocol that maps AI-generated logic to the engine's deterministic networking layer.
- The integration of DeepSeek V4 Pro and Kimi K2.7 is facilitated via a unified middleware layer that standardizes prompt formatting and context window management across different model providers.
- The visual-to-Lua transition uses a transformer-based translation model trained on millions of pairs of Eggy Code blocks and their corresponding Lua script equivalents.
🔮 Future ImplicationsAI analysis grounded in cited sources
⏳ Timeline
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: 36氪 ↗
