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
NVlabs/ToolOrchestra
ToolOrchestra is a research code release from NVIDIA Labs and the University of Hong Kong that trains compact orchestrator models to solve complex agentic tasks by delegating work to other tools and models instead of doing everything in one large model. The orchestrator alternates between reasoning steps and tool calls over multiple turns. Its tool set spans basic utilities such as web search and a code interpreter, specialized LLMs for coding and mathematics, and generalist frontier models, so the orchestrator learns which resource to invoke for each sub-problem and when a cheaper option suffices. Training is done with end-to-end reinforcement learning that combines three reward signals: outcome correctness, efficiency in terms of cost and tokens, and preference alignment. To supply enough training signal, the authors built an automatic pipeline that synthesizes environments and tool-call tasks at scale, published as the ToolScale dataset. The resulting Orchestrator-8B checkpoint is released on Hugging Face together with the training code, data generation pipeline and evaluation scripts. Reported results place the 8B model ahead of much larger frontier models on Humanity's Last Exam, tau2-Bench and FRAMES while using a fraction of the inference budget, and the project reports a top placement on the GAIA benchmark. The repository is aimed at researchers working on agentic reinforcement learning, tool-use routing and cost-aware model cascades, and at practitioners who want a small local router in front of expensive APIs. It is Python-based and ships the paper, project page and model links alongside reproduction instructions.
https://github.com/NVlabs/ToolOrchestra
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.