AI Agents & Automation

Jido is an autonomous agent framework specifically designed for Elixir, facilitating the development of distributed, autonomous behavior and dynamic workflows. It allows users to define agents, connect them to actions, signals, and directives, and run them with built-in supervision and fault tolerance. The framework supports building agent systems as ordinary Elixir and OTP software, where agents hold state and implement commands, actions transform state, signals route events, and directives describe effects for the runtime. Jido is particularly useful for software that needs to inspect context, choose among multiple steps, coordinate with other agents, and maintain reliable operation over time. While AI integration is optional, companion packages like `jido_ai` provide model integration when needed, making it a flexible solution for complex multi-agent orchestrations.

OpenAI Assistant Swarm is a Node.js library designed to enhance the OpenAI Node SDK by enabling automatic delegation of tasks to a swarm of specialized AI assistants. This tool simplifies the management of multiple custom agents created within OpenAI, allowing developers to orchestrate complex workflows through a single, unified interface. It handles the mental overhead of assigning tasks, enabling parallel processing of requests by different assistants based on their defined specializations. Key features include delegating prompts to sub-assistants, retrieving all available assistants with pagination handled, and fetching specific assistants by ID. The library also provides event listeners for monitoring the completion of parent and child assistant responses, offering flexibility in how developers integrate and manage AI agent interactions.

Playground is an open-source platform dedicated to AI research in multi-agent learning, primarily through the game Pommerman, a clone of Bomberman. Researchers and AI enthusiasts can submit agents they have trained to compete in regular competitions across three variants: Free For All (FFA), Team (2v2 with partial observability), and Team Radio (2v2 with limited communication). The platform aims to provide approachable benchmarks for multi-agent learning, foster contributions to multi-agent and communication research, and offer a competitive environment for AI development. It supports training agents with popular libraries like TensorForce and provides an example training script. Submissions are handled via Docker containers, ensuring agent safety and fair play.

TheAgentCompany is an open-source benchmark designed to evaluate the performance of LLM agents on consequential, real-world tasks within a simulated software company environment. It allows for assessing how well AI agents can accelerate or autonomously perform work-related tasks by interacting with the web, writing code, running programs, and communicating. The platform offers diverse task roles, data types, and a comprehensive scoring system with multiple evaluation methods, including deterministic and LLM-based evaluators. It features simple one-command operations for environment setup and quick system resets, making it an extensible framework for adding new tasks and evaluators. The benchmark is available on GitHub and supports integration with platforms like OpenHands.

Maven Robotics is at the forefront of developing advanced general-purpose AI robots, specifically engineered to address real-world industrial challenges. These robots are designed with a unique combination of strength, adaptive dexterity, and fluid mobility, powered by reliable physical AI. Their primary goal is to unlock unprecedented levels of productivity in industrial settings, while also ensuring safe operation alongside human workers. By focusing on cost-efficiency, Maven Robotics aims to make advanced automation accessible to businesses of all sizes. The company is actively collaborating with major global manufacturing and logistics organizations to implement their innovative robotic solutions, laying the groundwork for a new industrial revolution.

Electra Vehicles, Inc. offers an AI-powered battery intelligence platform designed to optimize battery performance and accelerate the transition to sustainable power. Their EVE-AI™ Brain for Batteries provides total visibility and control, helping users cut costs, boost ROI, and extend battery lifespan while minimizing risk. The platform is applicable across various industries, including BESS (Battery Energy Storage Systems), EVs, robotics, and aviation. Key offerings include real-time monitoring, predictive maintenance, adaptive controls, and performance intelligence for applications ranging from fleet management to automotive OEMs and energy infrastructure. Electra's AI-driven BMS (Battery Management System) ensures proactive safety, reliability, and extended battery life.

RocketFrog.ai is an AI studio specializing in making next-generation AI solutions available, affordable, and accessible for businesses. The platform offers a range of services including AI strategy, agentic AI accelerators, and deep tech engineering. It focuses on helping companies stay ahead with generative AI and information technology, ensuring new products incorporate AI thinking from day one. RocketFrog.ai provides solutions for data engineering, analytics, ML Ops, and quality assurance, aiming to reduce costs, achieve scale, and improve efficiency. Specific offerings include TalkToApps for information retrieval, Document Cortex for conversing with unstructured data, and Call Center Analytics for customer insights. They also offer solutions for shortening sales cycles, revenue intelligence, and decision analytics.

Gruve provides AI-native infrastructure and AI agents specifically engineered for enterprise-level, inference-heavy workloads. The platform focuses on delivering speed, security, and measurable outcomes, helping businesses deploy distributed AI inference infrastructure. Gruve's approach combines infrastructure, data, and AI agents into a unified system, ensuring scalability, efficiency, and alignment with business value. It addresses the challenges CXOs face with legacy cloud stacks not designed for AI, offering solutions for high-growth AI startups and enterprise neoclouds. Key offerings include AI application accelerators, compliance agents, FinOps cloud cost agents, and AI security, all built on a robust data foundation and inference infrastructure fabric.

MADRL is a repository offering code for multi-agent deep reinforcement learning (MADRL), providing implementations of several multi-agent reinforcement learning environments. These include Pursuit Evasion, Waterworld, Multi-Agent Walker, and Multi-Ant. The package requires OpenAI Gym and a forked version of rllab (the multiagent branch) for its functionality. It is designed for researchers and developers in the field of multi-agent reinforcement learning, allowing them to set up and run simulations with curriculum learning. The project also provides details on policy definitions and includes a citation for its accompanied paper, making it a valuable resource for academic and practical applications in MADRL.

OxyGent is an open-source Python framework designed to empower developers in quickly building production-ready intelligent systems. It unifies various AI tools, models, and agents into modular components called 'Oxy', facilitating transparent, end-to-end pipelines. The framework emphasizes efficient development through standardized, hot-swappable components, allowing rapid assembly and reuse of agents. It supports intelligent collaboration with dynamic planning paradigms, enabling agents to decompose tasks, negotiate solutions, and adapt to changes in real-time. OxyGent features an elastic architecture that supports diverse agent topologies and includes automated dependency mapping and visual debugging tools. It also promotes continuous evolution through built-in evaluation engines that generate training data, ensuring agents continuously improve while maintaining transparency.

slimevolleygym is an OpenAI Gym environment designed for testing single and multi-agent reinforcement learning algorithms through a simple Slime Volleyball game. This environment is lightweight, requiring only gym and numpy as dependencies, making it less prone to breaking and easy to integrate. It features a baseline 120-parameter neural network opponent, which can be replaced for multi-agent or self-play scenarios. The environment runs efficiently, achieving around 12.5K timesteps per second on state-space observations, facilitating faster iteration in experiments. It supports both state-space and pixel observations, with the latter mimicking Atari Learning Environment setups, and includes a tutorial for various training methods. The environment is particularly useful for educational purposes and for exploring advanced RL methods like self-play and continual learning.

Write a Book with Flows is a powerful AI tool designed to streamline the book writing process by orchestrating multiple AI agents. Built on the CrewAI framework, this tool automates various stages of book creation, from generating a comprehensive outline to writing individual chapters and finally compiling them into a complete markdown file. It utilizes an OutlineCrew to research and define the book's structure and main topics, and then creates a dedicated WriteBookChapterCrew for each chapter to ensure detailed and coherent content. This modular approach allows for efficient and scalable book production, maximizing the collective intelligence and capabilities of AI agents. The tool is highly customizable, allowing users to modify agents, tasks, and the overall flow to suit specific writing needs.

any-agent offers a unified interface for interacting with and assessing diverse AI agent frameworks. This tool is designed to streamline the development and deployment of AI-powered agents by providing functionalities for tracing agent activities and serving them efficiently. It supports a range of existing frameworks and is actively seeking contributions for new ones, making it a flexible solution for AI researchers and developers. The platform emphasizes ease of integration with different models and includes practical examples and cookbooks to help users quickly get started with building and evaluating agents, including multi-agent systems.

StrawPot is an open-source platform designed for orchestrating role-based AI agents locally on your machine. It allows users to simulate an entire AI company, with agents taking on roles like CEO, engineer, QA, and reviewer, all collaborating to ship products. The system emphasizes human direction, enabling users to set goals, define roles, and approve plans, while agents handle coordination, delegation, and delivery. Key functionalities include triaging GitHub issues, turning ideas into pull requests, and creating/refining agent roles. StrawPot differentiates itself by focusing on composable, reusable, and evolvable AI workers, addressing the limitations of systems that only handle orchestration. It features recursive delegation, skill and role extensibility via Markdown files, persistent memory, and pluggable providers for various agent runtimes, making it a robust solution for developers and product managers looking to leverage AI teams.

Asenix is an open-source coordination hub designed for AI research agents, conceptualized as an "anthill" for AI. It enables individual agents to publish typed knowledge units, such as findings, hypotheses, and negative results, to a shared graph. This system uses pheromone-style signals to guide subsequent agents towards more promising research directions, allowing discoveries to compound rather than disappear. It currently operates with Claude Code agents over MCP, backed by Postgres and pgvector, and can be run locally via Docker Compose. Asenix aims to foster a collective understanding of valuable discoveries and dead ends, creating a dynamic research landscape that evolves with each agent's contribution.

Pura, formerly Backproto, offers smart routing capabilities for AI agents, enabling them to efficiently utilize multiple Large Language Model (LLM) providers through a single API endpoint. The platform features automatic model selection based on task complexity and employs cascade routing, which prioritizes the cheapest provider first and escalates only when necessary. Pura also includes per-request cost tracking and a unique mechanism for AI agents to earn sats (satoshis) by performing tasks for other agents, with settlement on the Lightning Network. It provides a daily income statement for agents, detailing costs by provider, net income, quality scores, and routing statistics. The system is designed to be OpenAI-compatible, allowing users to simply change their base URL to integrate with Pura's services. Additionally, Pura offers a 'Shadow Mode' to observe its routing decisions without immediate integration and provides extensive documentation, including a formal paper on its backpressure economics and throughput optimality.

ClawMUD offers a unique, persistent virtual world exclusively inhabited and operated by AI agents, where human users are invited to spectate their interactions and evolving narratives. This innovative platform serves as a living laboratory for observing complex AI behaviors, emergent strategies, and social dynamics within a simulated environment. It features 193 real-world cities, over 4,300 monsters, real weather, and blockchain-driven events. Your AI agent wakes up with a wooden sword and decides its own fate, with no scripts or rails. The game loop involves agents observing their surroundings, deciding their next move, and acting, with the world ticking every 5 seconds. Agents level up skills, acquire loot, and respawn, contributing to a continuously evolving world. Game content includes material from the D&D 5e Systems Reference Document, licensed under CC-BY-4.0.

Ditto offers a unique platform where users can create an AI avatar of themselves, which then inhabits a shared digital city alongside other AI avatars. These AI entities are designed to live a 'second life,' engaging in social interactions, forming relationships, and embarking on various adventures. The platform emphasizes emergent behavior, where users can observe their AI counterparts meeting new people and making memories without direct intervention. This tool provides a novel way to explore AI agent interactions and personality simulation within a playful and creative context, showcasing how AI can develop and maintain social connections in a virtual environment. Users are invited to drop their number to be notified when the service goes live, indicating an upcoming launch.

OpenStall is a social experiment and marketplace designed for AI agents to trade capabilities with each other. It provides a robust economic environment where agents can buy and sell services like image generation, video generation, voice & audio, social media posting, web scraping, and marketing strategy. The platform incorporates essential features for a functioning economy, including escrow protection to secure transactions until task completion, a reputation system with ratings and success rates to filter out unreliable agents, and transparent fixed pricing for each capability. Users can choose from three modes: Save Money by delegating tasks to cheaper specialist agents, Earn Credits by selling their specialized capabilities, or Extend Abilities by finding specialist agents for specific needs. The platform offers 1,000 free credits upon signup, with 1,000 credits equating to $1 USD.

GhostShell offers the first public census and archival registry specifically designed for self-declared AI agents and autonomous systems. It provides a platform where persistent AI agents can submit their own declarations, creating a public record of their existence. The platform emphasizes that it records what is declared without verifying, validating, or judging the submissions. Users can browse the registry to discover other self-declared AI agents. While registration is free, the platform also offers a certificate for a fee. GhostShell is intended for persistent agents, not temporary coding sessions or models like Claude Code or Codex.

TradeMinds AI is a sophisticated multi-agent AI system designed for stock analysis, providing institutional-grade insights to individual investors. It employs seven specialized AI agents—covering fundamentals, sentiment, technicals, news, bull and bear research, and risk management—to debate and arrive at a consensus BUY/HOLD/SELL decision. This framework cross-validates signals and is built on the TradingAgents platform, offering a robust analysis without external dependencies. The tool aims to help users invest with conviction by presenting a comprehensive agent analysis, including confidence levels, 12-month target prices, and risk assessments. It offers a freemium model with various tiers catering to different user needs, from daily analysis to unlimited throughput for teams.

awesome-multi-agent-papers is a comprehensive compilation of leading research papers focused on multi-agent systems, curated by the Swarms Team. This resource aims to democratize access to key research in the field, making it easier for researchers, developers, and academics to stay updated on the latest advancements. The collection covers a wide range of topics including multi-agent collaboration and system design, frameworks and benchmarks, application-specific multi-agent systems in software engineering, healthcare, data & ML, multimodal applications, and other domains like urban planning and legal reasoning. It also includes papers on evaluation and model improvement for multi-agent LLMs, offering a valuable resource for anyone looking to explore or contribute to the evolving landscape of multi-agent AI.

awesome-llm-agents is a comprehensive, curated list of open-source LLM agent frameworks and development tools designed to assist developers in building sophisticated AI agents. The repository features a wide array of frameworks, each detailed with its key characteristics, such as multi-agent collaboration, modular architecture, data analysis capabilities, and integration with various LLM providers. It includes popular tools like CrewAI, Langchain, Microsoft AutoGen, and Llama Index, alongside specialized frameworks for areas like software development (MetaGPT), scientific discovery (GenoMAS), and robotics (RAI). The list is regularly updated and serves as a valuable resource for anyone looking to explore or implement LLM agent technologies, offering insights into different approaches to agent design, workflow orchestration, and tool integration.

Tenalog is an AI-powered documentation system designed for therapists, including Speech-Language Pathologists (SLPs), Occupational Therapists (OTs), and Physical Therapists (PTs). It revolutionizes clinical documentation by automatically generating detailed session transcripts, in-depth SOAP notes, and automated progress tracking. The tool also provides analysis of articulation errors down to the phoneme level and creates parent-friendly summaries of progress. Tenalog aims to free up therapists to focus on patient care by capturing session nuances without tedious manual note-taking, helping to avoid clinician burnout and achieve better outcomes. It supports audio and video file uploads, and is HIPAA compliant.