Article reprint source: NingNing
Spent some time researching the differences between the current version of AI16Z's AI Agent framework Eliza and the V2 version revealed by Shawn.
Before comparing the two, let's first understand the basic architecture of the current version and V2 version.
As shown in Figure 1, the current version of the Eliza framework consists of the Interface Layer, Core Layer, and Model Layer.
Interface Layer
Responsible for interactions with external platforms
Supports multiple platforms including Discord, Twitter, Telegram, etc.
Provides API interfaces for other systems to call
Core Layer
Memory system: Provides retrievable storage capabilities and document processing abilities
Behavior system: Manages custom behavior and action execution
Configuration system: Manages configuration through .env and elizaConfig.yaml
Model Layer
Supports various large language models, including OpenAI, Anthropic, Llama, Grok, etc.
Provides a unified model interface abstraction
Layers interact through standard interfaces, supporting high modularity and scalability.
On the 15th, Shawn revealed that they are developing Eliza v2, which is expected to launch an alpha version in 2 to 3 weeks.
Eliza v2 will unify the message bus and simplify the client; unified Agent wallet; model system adopts registry and overlay modes; enhances scalable general core framework; updates community plugins; achieves 100% test coverage.
The architectural design diagram of the V2 version is speculated as follows:
Let me explain the main improvements of Eliza v2 and its position in the architecture:
Improvements in the Interface Layer: Unified message bus
Integrated the originally dispersed client into a unified message processing system
Each platform (Discord/Twitter, etc.) becomes an adapter for the message bus
Simplified client development and maintenance work
Provides a more consistent message processing experience
Improvements in the Core Layer: Unified proxy wallet
Integrated the originally dispersed wallet configurations
Provides unified multi-chain wallet management
Simplifies cross-chain operations
Scalable core framework: Redesigned core interfaces to be more general
Enhances the framework's scalability
Better plugin integration mechanism
Community plugin updates: Improved plugin system
Improves plugin management
100% test coverage guarantees quality
Improvements in the Model Layer: Introduces a model registry mechanism
Supports dynamic registration and overlay of models
Divides models into two major categories: LLMs and Embeddings
More flexible model invocation methods
The architecture of this new version is more modular and unified, with clearer interactions between different components, providing a better foundation for future expansions.
Finally, from an investment perspective, Eliza v2 will bring new growth opportunities for feature extension plugins like AgentTank.