Open Source Frameworks
Discover 104 open source Framework repositories from GitHub, each with an analysis of what it does, key features, use cases and alternatives. Framework projects here are most often combined with Python, LLM and AI Agents. Last updated October 4, 2026.
104 repositories · updated October 4, 2026
RouteLLM: Route LLM Requests to Balance Cost and Quality
RouteLLM is a Python framework for routing prompts between stronger, costlier models and cheaper alternatives. It helps teams manage the cost-quality tradeoff, serve requests through an OpenAI-compatible interface, and evaluate routing strategies.
LightLLM: Serve Large Language Models with a Python Framework
LightLLM is a Python framework for LLM inference and serving, designed for scalable deployment and fast generation. It suits teams operating model-serving systems and researchers building on inference components.
LazyLLM: Build Multi-Agent LLM Applications
LazyLLM is a Python framework for assembling, testing, and deploying multi-agent LLM applications with low-code workflows. It suits developers who want to combine models, tools, and RAG components while iterating across local or online services.
ChatArena: Build Multi-Agent Language Game Environments
ChatArena is a Python framework for running language games with multiple LLM agents. It suits researchers and developers studying agent interaction, collaboration, and social behavior, but the project was deprecated in August 2025 and is no longer supported.
Agentarium: Build and Orchestrate AI Agent Simulations
Agentarium is a Python framework for creating AI agents that interact, take context-based actions, and retain memories. Use it to prototype multi-agent scenarios, add custom actions, and save agent states for repeatable experiments.
lighteval: Evaluate Language Models Across Backends
Lighteval is a Python toolkit for running LLM evaluations across local models and remote inference backends. It combines a broad task catalog with custom metrics and detailed sample-level results for teams comparing or debugging model performance.
agent-native: Build Agentic Apps with Shared UI and Actions
Agent-Native is a TypeScript framework for building applications where people and AI agents work through the same actions and shared data. It suits teams creating agent-powered tools that need an interactive UI, permissions, and workflows beyond chat.
EasyJailbreak: Build and Evaluate LLM Jailbreak Attacks
EasyJailbreak is a Python framework for assembling and testing jailbreak methods against language models. It suits researchers and developers who need reusable attack components and a structured way to evaluate model responses.
guardrails: Validate LLM Inputs and Outputs
Guardrails is a Python framework for checking LLM inputs and outputs against configurable validators and for generating structured data. It suits teams building AI applications that need validation and error handling around model responses.
physicsnemo: Build and Train Physics AI Models
NVIDIA PhysicsNeMo is a PyTorch framework for building and training machine-learning models for physics and engineering. It combines reusable model components with end-to-end recipes for scientific data and workloads.
Lean: Build and Run Algorithmic Trading Strategies
QuantConnect Lean is an event-driven engine for developing, backtesting, and live trading algorithmic strategies across financial markets. It suits quantitative developers who want a modular C# engine with Python support and local or cloud-assisted workflows.
tensorrec: Build Custom Recommendation Systems with TensorFlow
TensorRec is a Python framework for building recommendation systems with TensorFlow. It combines user and item features with interaction data, while letting developers customize representation and loss functions. The project is not under active development.