# Summary of code walkthrough of Hyperfy v2
## Summary (Artisinal)
> **Model Used:** maximus
> **Prompt:** Summarize and takeaways
> **Link to Source:** [Twitter](https://x.com/i/broadcasts/1BdxYEygmLgxX)
> **Summary:** Ash from the core development team sharing initial release of Hyperfy v2 code
### Informal review of the code released today:
* As background: Hyperfy is an engine/platform for immersive and interactive apps
* Apps behave like browser plugins sandboxed from each other using are secure SCS (private compartments)
* Apps will be configurable and the code will also be editable
* Is a refined version of javascript to allow simpler coding without relying on more complex frameworks
* Goal is to be able to create whole self-contained worlds and modular apps that can be readily shared and used
* As easy to start as build apps should start with a 3d model, attach scripts and configuration on it
### Currently 2 main git repos opened up by Hyperfy
* Reference Engine (reimagining of the v1)
* There’s a post that describes things that can be done such as mad instancing, physics
* Coding in browser enabled here; eliminating the need to do coding/development like it was needed for v1
* Intended as a test build not fully hostable (ex. Uses Firebolt)
* Hyperfy Self-hostable Engine (v2 repo)
* Crafted to be be self-hostable
* Does not have all the features for Reference Engine (but will be updated soon to have everything the reference engine has)
### Questions: starting from 7:28pm**
* Can we back up worlds? Yes, technically create fully backed up worlds embedded in NFT
* 7:32 Will all v1 emotes work in v2… for example you can create an object and combine with an emote
* 7:36 = For someone who doesn’t want to host what options will they have? Eventually there will be packaged providers that can provide the back-end
* Can we use blockchain? Yes, will be chain agnostic, chains will be supported by apps
* 8:29pm Q: Pathfinding for agents? Yes, better than v1. Can use pathing as well as “follow” commands
* 8:33 Q: Avatar backpacks? Yes, planned to improve data portability
* 8:36 Q: Policing malicious apps? Users will be responsible for their pages, no central authority since things are self hosted
* 8:38 Q: App ratings? We thought of it but it isn’t necessary yet, not our place to police/host a market place
* Will streaming work from v2? A: Will require porting apps from v1 primarily supporting HLS and m3u8
* Does flying work? A: Yes PhysX integrated can really fly!
* How big can a world be? A: Very big, limited by javascript floating point
* How to control avatars to sit or aim? 1) That is an object app with an emote bound to it + 2) A core app is needed to change perspectives to do things like FPS perspectives
### Ash spinning up a of v2 repo pull
* Meant to be simpl npm install with very basic dependencies (meant for easy maintenance)
* Basically a git clone npm etc and it spins up an Sqlite database to hold data
* Intent is to run on low compute servers such as 512mb (easy self hosting)
* 8:16 talking about how engine in it’s purest form is basically a javascript handler for DOM/colliders/LOD to create immersive worlds rich with plugins
### Key Next Steps (7:39pm – 7:51 + 8:03)
* 1. Persistence for worlds to be fully self-hostable
* 2. Identity service layer will be created (so people can save their Avatar accessible across worlds)
* 3. Code editor will be moved into
* 4. World import /export capability (mirror)
* 5. Versioning app scripts (making it so apps work forever)
* 6. Docs plzzz
* 7: 40 = Requests to the public
* Try it out, give feedback
* If you are keen and understand how the 3js in working :: We have apps from the reference engine AND v1 that needs things ported over to the core engine
### Far Future:
* Backpack files across worlds
* Let people whitelist certain types of apps to be usable in app
## Summary (LLM)
> **Model Used:** OpenAi o3
> **Prompt:** Summarize transcript into Platform | Code | Features | Questions and Next Steps
> **Link to Source:** [Twitter](https://x.com/i/broadcasts/1BdxYEygmLgxX)
> **Summary:** Ash from the core development team sharing initial release of Hyperfy v2 code
### Overview of the Hyperfy Platform
* Primary Objective: To simplify virtual world-building and interactivity, targeting accessibility for both developers and non-developers.
* Core Innovations:
* In-browser coding capability, eliminating SDK installation requirements.
* Persistent, self-hostable worlds with drag-and-drop features.
* Export functionality for interoperable assets in a single file.
* Introduction of cryptographically signed apps for secure sharing and integration.
* Streamlined resource requirements allowing for cost-efficient hosting (e.g., servers with 512MB RAM).
* Potential Applications:
* Embedding apps in NFTs for innovative use cases.
* Creating open ecosystems akin to Minecraft servers, with enhanced portability.
* Modular marketplace concept for assets and apps.
### Code Development Experience* Development Flow:
* Developers can start with basic commands like npm install to spin up a local environment.
* World-building involves creating directories containing SQLite databases and asset folders.
* Drag-and-drop mechanics streamline 3D asset integration, with minimal dependencies for faster deployment and testing.
* Flexibility for Customization:
* Developers can embed specific functionalities (e.g., emotes, interactive objects) using pre-set configurations.
* Physics and interactive object creation are simplified through customizable scripts.
* Ease of Deployment:
* Worlds are scalable and can support hundreds of concurrent users.
* Features include modularity for switching between world configurations and seamless server transitions.
### Showcase of Key Features
* Interactive World Dynamics:
* Real-time collaboration allows users to edit or modify elements in live environments.
* Apps can take over user controls for immersive experiences (e.g., sitting animations, custom interactions).
* Enhanced App Integration:
* Apps function as modular elements, providing flexibility for adding functionality like physics-based objects or Blockchain integrations.
* Sandbox environments isolate app behaviors, ensuring world stability.
* Performance Highlights:
* Efficient handling of assets and physics-based simulations (e.g., multi-player synced environments).
* Supports lightweight hosting solutions, enabling low-cost scalability.
* Customization Capabilities:
* Developers can create unique controllers for specific use cases, such as first-person or racing simulations.
* Compatibility with existing assets and flexibility to integrate advanced features like AI agents or live media streaming.
### Open Questions
* How will media streaming evolve in V2 to accommodate user-hosted solutions?
* What enhancements are planned for avatar interaction and customization (e.g., integrating dynamic emotes)?
* How will Hyperfy encourage third-party contributions while maintaining platform integrity?
### Next Steps
* Finalize integration of critical features like persistence, app import/export, and modular configurations.
* Establish tools and support for developers, particularly in areas like identity management and advanced scripting.
* Explore collaborative opportunities for creating ecosystems, such as marketplaces or pre-configured app bundles.
* Focus on user feedback to refine the platform’s usability and scalability.
{%hackmd Syq4xkPwkg %}