aser 0.1.8

An AI Agent

Aser

Aser is a minimalist, modular, and versatile AI agent framework. You can assemble an agent with just a few lines of code.

Website | Documentation | Get Support | 中文

Installation

Install from pypi:

pip install aser

Or clone the repository:

git clone https://github.com/AmeNetwork/aser.git
cd aser
pip install -r requirements.txt

Set up environment variables

Please refer to .env.example file, and create a .env file with your own settings. You don't need to configure all environment variables, just select the ones you use.

.env file example:

#MODEL
MODEL_BASE_URL=<your model base url>
MODEL_KEY=<your model key>

Usage

#Basic
from aser.agent import Agent
agent=Agent(name="aser agent",model="gpt-4.1-mini")
response=agent.chat("what's bitcoin?")
print(response)

# Full configuration
aser = Agent(
    name="aser",
    model="gpt-4o-mini",
    tools=[web3bio, exa],
    knowledge=knowledge,
    memory=memory,
    chat2web3=[connector],
    mcp=[price],
    trace=trace
)

Get Started

If you clone the project source code, before running the examples, please run pip install -e . in the root directory, which allows Python to find and import the aser module from the local source code. If you install it via pip install aser , you can run the examples directly.

Beginner:

• Aser Agent: Your First AI Agent
• Model Config: Customize the LLM configuration
• Memory: Build an agent with memory storage
• RAG: Build an agent with knowledge retrieval
• Tools: Build an agent with tools
• Toolkits: Use built-in toolkits
• Trace: Build an agent with tracing
• API: Build an agent with API server
• CLI: Interact with the agent using the CLI
• Discord: Build an agent with Discord client
• Telegram: Build an agent with Telegram client
• Farcaster: Build an agent with Farcaster client

Intermediate:

• CoT: Chain of Thought
• MCP: Model Context Protocol
• Workflow: Custom Agent Workflows
• Evaluation: Evaluate an AI Agent
• Router Multi-Agent: Multiple agents distribute tasks based on routing
• Sequential Multi-Agent: Multiple agents work sequentially
• Parallel Multi-Agent: Multiple agents work simultaneously
• Reactive Multi-Agent: Multiple agents respond to changes
• Hierarchical Multi-Agent: Multiple agents work at different levels
• Agent UI: Interact with the agent through the UI

Advanced:

• MSCP: Model Smart Contract Protocol
• A2Aser: Integrate Google Agent2Agent (A2A) Protocol
• A2A Client: Agent to Agent Client

Experiments:

• ERC8004 Server Agent: Server Agent Offers services and executes tasks
• ERC8004 Identity: Identity registration is performed using the MSCP ERC8004 Connector