nowfound

AI · October 28, 2026

A Lego-like composable Framework for RSI

GEAR: An Open General Evolutionary Architecture for Recursive Self-Improvement of LLM Agent Harness. An open framework for harness evolution, reproducible evaluation, and learning across tasks.

A Lego-like composable Framework for RSI launched on October 28, 2026 with 18 votes, #841 of 4,984 launches that month and more than 90% of that year's launches.

Visit rsi-gear.github.io ↗Alternativestop 17% of October 2026

In plain words

GEAR is an open framework for building and evolving large language model agents through a two-level learning process. An inner loop refines an agent's reasoning and actions on individual tasks using feedback, while an outer loop identifies patterns across multiple tasks and proposes improvements that enhance overall performance. The framework enables reproducible evaluation and learning, with composable components that can be optimized independently, making it suitable for researchers and developers working on recursive self-improvement in AI agents.

written from the facts on this page · October 2026

From the sources

GEAR organizes recursive self-improvement as a two-level meta-learning process. In the inner loop , a rollout agent uses task feedback to refine its reasoning, planning, and actions within a task. In the outer loop , a Meta Agent identifies common patterns across task experiences, including recurring failures and effective strategies, and proposes reusable improvements to enhance cross-task performance. GEAR evaluates and promotes these changes for subsequent tasks. The two loops can be optimized independently and are connected by experience. GEAR organizes evolutionary search into replaceable algorithm components, providing a common framework for studying different search strategies.…from rsi-gear.github.io

Does a similar job

all alternatives →
  • Evolving Agents Framework2025 · github.com · ▲139

    Hey HN, I've been working on an open-source framework for creating AI agents that evolve, communicate, and collaborate to solve complex tasks. The Evolving Agents Framework allows agents to: Reuse, evolve, or create new agents dynamically based on semantic similarity Communicate and delegate tasks to other specialized agents Continuously improve by learning from past executions Define workflows in YAML, making it easy to orchestrate agent interactions Search for relevant tools and agents using OpenAI embeddings Support multiple AI frameworks (BeeAI, etc.) Current Status & Roadmap This is…

  • agentmash18d ago · agentmash.dev · ▲142

    open-source multiplayer AI for teams' coding agents

  • OS
  • Praex30d ago · praex.ai · ▲68

    The coding agent that runs on its own AI models

  • OpenAI Gym2016 · ▲147

    Open source interface to reinforcement learning tasks

  • Agent-harness-kit scaffolding for multi-agent workflowsMay 2026 · ahk.cardor.dev · ▲82

    A collection of open-source TypeScript tools by @enmanuelmag. heimdall-mcp for MCP proxy telemetry. agent-harness-kit for multi-agent orchestration.

More ai this month

the category →

Launched alongside, October 2026

the whole month →