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
langfengQ/verl-agent
verl-agent is an advanced extension of the veRL framework, specifically designed to train large language model (LLM) and vision-language model (VLM) agents using reinforcement learning (RL). It addresses the challenges of long-horizon, multi-turn RL tasks by handling each interaction step independently rather than concatenating entire interaction histories. This approach allows for fully customizable input structures at each step, making the system highly scalable for complex tasks requiring many sequential steps, such as those found in ALFWorld which can demand up to 50 steps for task completion. The project supports a wide range of RL algorithms, including the novel Group-in-Group Policy Optimization (GiGPO), GRPO, PPO, DAPO, RLOO, and REINFORCE++, among others. It also offers a rich suite of environments for training agents, spanning text-based and visual tasks. Supported environments include ALFWorld, Sokoban, Gym Cards, WebShop, and an experimental AppWorld, providing diverse scenarios for agent development. verl-agent features parallelized gym-style environments and introduces group environments where multiple rollouts share the same initial state, facilitating efficient training for group-based RL algorithms. It supports various large language models such as Qwen3, Qwen2.5, Qwen2.5-VL, and LLaMA3.1, and includes support for LoRA fine-tuning to reduce computational costs, enabling training of large models with limited GPU resources. The system supports multi-modal agents capable of reasoning with both text and images, expanding its applicability to vision-language tasks. Additionally, it includes a prompt-based agent leveraging GPT-4o for enhanced interaction capabilities. Overall, verl-agent is a comprehensive and scalable platform for training sophisticated LLM and VLM agents through reinforcement learning, equipped with state-of-the-art algorithms, flexible input handling, and a variety of challenging environments.
https://github.com/langfengQ/verl-agent
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.
aiwaves-cn/agents
Agents 2.0 is an open-source framework that enables the training and self-evolution of autonomous language agents through symbolic learning inspired by neural network methodologies.
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.