headroomlabs-ai/headroom
Headroom is a context compression layer for AI agents, reducing token usage by 60-95% across tool outputs, logs, RAG chunks, files, and conversation history, while maintaining accuracy.
Awesome AI Agents › End-to-End Optimizers
aiwaves-cn/agents
The project "Agents 2.0" is an open-source framework designed for creating data-centric, self-evolving autonomous language agents. It introduces a novel approach called agent symbolic learning, which draws inspiration from the connectionist learning procedures used in neural networks. The framework establishes an analogy between language agents and neural nets, where the agent pipeline corresponds to the computational graph of a neural net, nodes in the pipeline correspond to layers, and prompts and tools correspond to weights. This analogy allows the implementation of key neural network training components such as backward propagation and gradient-based weight updates in the context of language agent training. The framework operates by conducting a forward pass where the agent executes tasks and stores inputs, outputs, prompts, and tool usage in a trajectory. A prompt-based loss function then evaluates the outcome, producing a language loss. This loss is back-propagated through the nodes, generating textual analyses and reflections termed language gradients. These gradients are used to update the symbolic components and the computational graph of the agent, enabling self-evolution and learning. The system also supports multi-agent optimization by treating nodes as different agents or allowing multiple agents to act within a single node. Agents 2.0 is designed to facilitate the training and evaluation of autonomous language agents with a focus on symbolic learning and self-improvement. It provides a structured and systematic framework that leverages language-based loss functions and gradient updates to enhance agent performance over time. The project includes comprehensive documentation, installation instructions, and is actively maintained with recent updates. It is licensed under Apache 2.0 and encourages contributions and citations for academic use. Overall, Agents 2.0 represents a significant advancement in the development of autonomous language agents by integrating symbolic learning techniques with established neural network training methodologies, enabling agents to evolve and improve autonomously through language-based feedback mechanisms.
https://github.com/aiwaves-cn/agents
Also in End-to-End Optimizers
topoteretes/cognee
Cognee is an open-source project that provides scalable and modular memory solutions for AI agents using ECL pipelines, enabling efficient data ingestion, knowledge graph generation, and query capabilities to enhance AI performance and reduce hallucinations.
letta-ai/letta
Letta is an open-source framework for building stateful AI agents with advanced reasoning, transparent long-term memory, and context management, supporting multiple LLM backends and providing a graphical development environment.
langchain-ai/opengpts
OpenGPTs is an open-source platform that replicates and extends OpenAI's GPTs and Assistants API, offering customizable cognitive architectures, extensive tooling, and support for multiple language models and vector databases.
getzep/zep
Zep is an AI memory foundation that uses a temporal Knowledge Graph to enable continuous learning and personalized experiences for AI agents by efficiently managing and retrieving contextual chat and data artifacts.
algorithmicsuperintelligence/optillm
OptiLLM is an optimizing inference proxy that significantly improves the accuracy and performance of large language models on reasoning tasks without requiring any model training, by applying over 20 advanced optimization techniques during inference.
mll-lab-nu/RAGEN
A reinforcement learning framework for training multi-turn reasoning LLM agents on the StarPO algorithm, shipping ten Gym-compatible environments plus diagnostics for echo-trap and reasoning-collapse failure modes.
langfengQ/verl-agent
verl-agent is an extension of veRL designed for scalable training of large language and vision-language model agents via reinforcement learning, featuring customizable input structures, diverse RL algorithms including GiGPO, and support for multi-modal environments.