Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
A transparent, minimal, and hackable agent framework. ~300 lines of readable code. Full control, no magic.
| Date | Stars |
|---|---|
| 2026-07-31 | 453 |
| 2026-08-05 | 453 |
| 2026-08-06 | 453 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# AgentSilex
[](https://badge.fury.io/py/agentsilex)
[](https://opensource.org/licenses/MIT)
[](https://www.python.org/downloads/)
[](https://pepy.tech/project/agentsilex)
A transparent, minimal, and hackable agent framework for developers who want full control.
**Read the entire codebase in one sitting. Understand exactly how your agents work. No magic, no hidden complexity.**
## Why AgentSilex?
While large agent frameworks offer extensive features, they often become black boxes that are hard to understand, customize, or debug. AgentSilex takes a different approach:
- **Transparent**: Every line of code is readable and understandable. No magic, no hidden complexity.
- **Minimal**: Core implementation in ~300 lines. You can read the entire codebase in one sitting.
- **Hackable**: Designed for modification. Fork it, customize it, make it yours.
- **Universal LLM Support**: Built on LiteLLM, seamlessly switch between 100+ models - OpenAI, Anthropic, Google Gemini, DeepSeek, Azure, Mistral, local LLMs, and more. Change providers with one line of code.
- **Educational**: Perfect for learning how agents actually work under the hood.
## Who is this for?
- **Companies** who need a customizable foundation for their agent systems
- **Developers** who want to understand agent internals, not just use them
- **Educators** teaching AI agent concepts
- **Researchers** prototyping new agent architectures
## Demo
*Real-time streaming response demonstration - see how AgentSilex processes queries and streams results*
## Installation
```bash
pip install agentsilex
```
Or with uv:
```bash
uv add agentsilex
```
## Quick Start
```python
from agentsilex import Agent, Runner, Session, tool
# Define a simple tool
@tool
def get_weather(city: str) -> str:
"""Get weather information for a city."""
# In production, this would call a real weather API
return "SUNNY"
# Create an agent with the weather tool
agent = Agent(
name="Weather Assistant",
model="gemini/gemini-2.0-flash", # Switch models: openai/gpt-4, anthropic/claude-3-5-sonnet, deepseek/deepseek-chat, et al.
instructions="Help users find weather information using the available tools.",
tools=[get_weather]
)
# Create a session to track conversation history
session = Session()
# Run the agent with a user query
runner = Runner(session)
result = runner.run(agent, "What's the weather in Monte Cristo?")
print(result.final_output)
# Output: "The weather in Monte Cristo is SUNNY."
```
**That's it!** In just 20 lines, you have a working agent that:
- ✅ Uses any LLM (OpenAI, Anthropic, Google, DeepSeek, etc.)
- ✅ Calls tools to get information
- ✅ Maintains conversation history
- ✅ Returns natural language responses
## Multi-Agent Example
AgentSilex supports intelligent agent handoffs, allowing a main agent to route requests to specialized sub-agents:
```python
from agentsilex import Agent, Runner, Session, tool
# Specialized weather agent with tools
@tool
def get_weather(city: str) -> str:
"""Get weather information for a city."""
return "SUNNY"
weather_agent = Agent(
name="Weather Agent",
model="openai/gpt-4o-mini",
instructions="Help users find weather information using tools",
tools=[get_weather],
)
# Specialized FAQ agent
faq_agent = Agent(
name="FAQ Agent",
model="openai/gpt-4o-mini",
instructions="Answer frequently asked questions about our products",
)
# Main orchestrator agent
main_agent = Agent(
name="Main Agent",
model="openai/gpt-4o-mini",
instructions="Route user questions to the appropriate specialist agent",
handoffs=[weather_agent, faq_agent],
)
# Execute multi-agent workflow
sessioExcerpt of 7,101 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:3e18d75206b9058b, topic:ai-agent, topic:agent-framework, desc:agent framework