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
gepa-ai/gepa
GEPA (Genetic-Pareto) is a Python framework for optimizing systems that have textual parameters against an arbitrary evaluation metric: prompts, code, agent architectures, scheduling policies, even vector graphics. Unlike reinforcement-learning or gradient-based optimizers that reduce an execution trace to a single scalar reward, GEPA has an LLM read the full trace -- error messages, profiling output, reasoning logs -- to diagnose why a candidate failed and propose a targeted textual mutation. Its loop selects a candidate from the current Pareto frontier (the set of candidates each excelling on different task subsets), executes it on a minibatch while capturing traces, has a reflection model diagnose the failure and propose an improved candidate, and accepts it into the pool if it advances the frontier; GEPA also supports "system-aware merge," combining the complementary strengths of two Pareto-optimal candidates. The core abstraction behind feeding diagnostic information back into reflection is called Actionable Side Information, described by the authors as a textual analogue of a gradient. Reported results include beating Claude Opus 4.1 at lower cost using open-source models plus GEPA, running much faster than GRPO-style reinforcement learning, lifting ARC-AGI agent accuracy substantially through automated agent-architecture discovery, and production use at companies including Shopify, Databricks and OpenAI. GEPA integrates with DSPy as dspy.GEPA for prompt optimization inside AI pipelines, and exposes a general optimize_anything API that only requires an evaluator function and a target artifact. It also ships as an installable Agent Skill so coding agents such as Claude Code can drive the optimizer directly. The project is MIT licensed, based on a 2025 research paper.
https://github.com/gepa-ai/gepa
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.