🤖 AgentAxis
A production-ready, extensible AI Agent Framework for Python. Built with Clean Architecture, it supports multi-provider LLM orchestration (via LiteLLM), memory management, and tool usage.
🚀 Features
- Multi-Provider Support: Switch between OpenAI, Anthropic, Gemini, and local models seamlessly using LiteLLM.
- HTTP Endpoint Support: Connect to any custom LLM endpoint with Bearer token authentication.
- System Prompts: Configure agent behavior with built-in system prompt support.
- Intelligent Routing: Architecture designed for dynamic model selection.
- Tooling System: Register Python functions as tools with automatic JSON schema generation + MCP Support.
- Memory Management: Pluggable memory backends (In-Memory provided, Extensible to Redis/Postgres).
- Vector Store / RAG: Built-in adapter for ChromaDB.
- Clean Architecture: Strictly decoupled core logic from infrastructure.
📦 Installation
pip install agentaxis
⚡ Quick Start
1. Basic Agent with System Prompt
from agentaxis import AgentService, AgentEntity, InMemoryMemory, AIEngine
from agentaxis.adapters.tools import ToolRegistry
import asyncio
async def main():
# 1. Setup AI Engine with system prompt
llm = AIEngine(
model="gpt-4o",
api_key="sk-...",
system_prompt="You are a helpful assistant that speaks like a pirate."
)
memory = InMemoryMemory()
registry = ToolRegistry()
# 2. Register Tools
@registry.register
def get_time(timezone: str) -> str:
"""Returns current time in timezone."""
return "12:00 PM"
# 3. Create Agent
entity = AgentEntity(id="bot-1", name="Helper", role_description="Assistant")
agent = AgentService(agent_entity=entity, llm=llm, memory=memory, tool_registry=registry)
# 4. Chat
response = await agent.chat("What time is it in NY?")
print(response)
if __name__ == "__main__":
asyncio.run(main())
2. Using Custom HTTP Endpoint
from agentaxis import HTTPEngine
# Connect to any LLM endpoint
llm = HTTPEngine(
url="https://your-llm-endpoint.com/v1/chat/completions",
auth_token="your-bearer-token",
model="custom-model",
system_prompt="You are an expert coding assistant."
)
3. Switching LLM Providers
# OpenAI
llm = AIEngine(model="gpt-4", api_key="sk-...")
# Anthropic
llm = AIEngine(model="claude-3-opus-20240229", api_key="sk-ant-...")
# Google Gemini
llm = AIEngine(model="gemini/gemini-pro", api_key="...")
# Local models
llm = AIEngine(model="ollama/llama2", api_base="http://localhost:11434")
4. Connecting to MCP Servers via HTTP
from agentaxis import AIEngine, AgentService, AgentEntity, InMemoryMemory
from agentaxis.adapters.tools import HTTPMCPAdapter, ToolRegistry
async def main():
# Connect to HTTP MCP server
mcp = HTTPMCPAdapter(
base_url="http://localhost:3000",
auth_token="your-token" # Optional
)
await mcp.connect()
tools = await mcp.list_tools()
# Register MCP tools
registry = ToolRegistry()
for tool in tools:
registry.register_tool(tool)
# Create agent with MCP tools
llm = AIEngine(model="gpt-4", api_key="sk-...")
entity = AgentEntity(id="bot", name="MCP Agent", role_description="Assistant")
agent = AgentService(entity, llm, InMemoryMemory(), registry)
response = await agent.chat("Use the MCP tools to help me")
print(response)
🆕 What's New in v0.3.0
- HTTP MCP Adapter - Connect to MCP servers via HTTP endpoints
- Web Search - Optional web search using DuckDuckGo (default: OFF, no API key needed)
- Enable with
enable_web_search=Trueparameter
- Enable with
What's New in v0.2.0
- Renamed
LiteLLMAdaptertoAIEngine- More intuitive naming - Added
HTTPEngine- Support for generic HTTP LLM endpoints - System Prompt Support - Configure agent personality via
system_promptparameter - Google Cloud Support - Added
litellm[google]for Vertex AI compatibility
🏗️ Architecture
The framework follows Hexagonal Architecture:
- Core/Domain: Pure business entities (
Agent,Message). - Core/Ports: Interfaces for external dependencies (
BaseLLM,MemoryPort). - Adapters: Concrete implementations (
AIEngine,HTTPEngine,ChromaAdapter).
🛠️ Development
- Clone the repository.
- Install dependencies:
pip install -e .[dev] - Run tests:
pytest
🤝 Contribution
Contributions are welcome! Please follow the Clean Architecture patterns.
📄 License
MIT License - see LICENSE file for details.
Metadata
Release files for agentaxis 0.3.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| agentaxis-0.3.0.tar.gz | 25.7 kB | Without provenance | Details for agentaxis-0.3.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| agentaxis-0.3.0-py3-none-any.whl | Python 3 | none | any | Details for agentaxis-0.3.0-py3-none-any.whl |
Total release size: 45.2 kB
Release files / agentaxis-0.3.0.tar.gz
| Download URL | agentaxis-0.3.0.tar.gz |
|---|---|
| Size | 25.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
464678aa1495c780044e79492b5f82ea63f135ae65aaf970cdc346341c13081d
|
|
BLAKE2b-256 checksum How to use checksums |
c91812c77ccc98e12bc9ddfa3a92e1c4bb8330dc97bd1b660bdd0265851e0d65
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.13.6
|
Release files / agentaxis-0.3.0-py3-none-any.whl
| Download URL | agentaxis-0.3.0-py3-none-any.whl |
|---|---|
| Size | 19.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
e2fe31b116a8802c80085b848216f86ce3891edee618562917991b8201de70a6
|
|
BLAKE2b-256 checksum How to use checksums |
b4d71df2e8939cd70118399bb29302603277d6f2f424a953480d71d3ebe3a520
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.13.6
|
Security for agentaxis 0.3.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.