14 dépôts
Systems for managing state, memory, and tool interactions across multi-step agent workflows.
Distinguishing note: Focuses on the orchestration of stateful agent cycles.
Explore 14 awesome GitHub repositories matching artificial intelligence & ml · Agent Task Orchestrators. Refine with filters or upvote what's useful.
This repository is a collection of guides, notebooks, and recipes for implementing advanced prompting techniques and workflow patterns with large language models. It serves as a prompt engineering guide, an evaluation suite for scoring prompt quality, and a framework for orchestrating agents and integrating external tools. The project provides implementation patterns for building applications with Claude, specifically focusing on coordinating multiple models to split complex tasks between high-reasoning and high-efficiency agents. It includes technical demonstrations for multimodal data proce
Implements patterns for coordinating multiple models to split complex tasks between high-reasoning and high-efficiency agents.
Phidata is an LLM agent framework and agentic workflow orchestrator used to build autonomous agents that integrate custom data, tools, and memory. It provides a production environment for serving these agents as services via APIs, utilizing server-sent events and websockets for real-time communication. The system distinguishes itself through a human-in-the-loop control layer that requires manual approval and administrative sign-off for specific tool executions. It also implements a multi-tenant AI infrastructure that uses token-based roles to ensure data isolation between different tenants.
Orchestrates state, memory, and tool interactions across complex multi-step agent workflows and scheduled tasks.
Picoclaw is a lightweight framework designed for the deployment and orchestration of autonomous software agents. It functions as a cross-platform runtime that packages the entire system into a single self-contained binary, enabling native execution across diverse hardware architectures including RISC-V, ARM64, and x86_64. The platform is specifically optimized for resource-constrained environments, ensuring minimal startup times and a low memory footprint. The system distinguishes itself through intelligent task orchestration, which routes incoming requests to specific inference models based
Manages complex workflows by routing tasks to appropriate inference models based on difficulty to optimize performance.
This framework provides a development toolkit for building autonomous agents that utilize language models to solve complex, non-deterministic tasks. Its core design centers on a code-executing architecture where agents generate and run Python code snippets to perform logic, data manipulation, and tool interactions. By moving beyond structured data formats, the system enables agents to manage program flow and object state through iterative reasoning cycles. The project distinguishes itself through its focus on code-based agent implementation and secure execution environments. Developers can ch
Manages agent state, tool interactions, and memory across iterative cycles to solve complex requests.
Letta is a framework for building, deploying, and managing autonomous AI agents that maintain persistent state across long-term interactions. It provides a comprehensive suite of primitives for defining agents with configurable personas, modular memory blocks, and tool-use capabilities, enabling them to retain user preferences and conversation history over extended sessions. The platform distinguishes itself through its advanced memory management and orchestration capabilities. It allows agents to autonomously update their own memory, perform retrieval-augmented generation, and coordinate com
Manages state, memory, and tool interactions across parallel agent workflows for knowledge synthesis.
This project is a comprehensive framework for developing, orchestrating, and deploying autonomous agents. It provides a structured environment for building agents that utilize reasoning loops to perform multi-step tasks, manage state through graph-based workflows, and interact with external tools. By mapping unstructured model outputs into typed schemas, the framework ensures reliable integration with downstream application logic. The platform distinguishes itself through a focus on production-grade reliability and security. It incorporates hybrid memory systems that combine vector embeddings
Coordinates multi-step agent workflows and manages state transitions for scalable operations.
This project is a Rust-based AI agent framework and tool orchestrator that provides a command-line interface for interacting with large language models. It functions as an AI tool orchestrator that routes client requests to language servers and manages the planning and handoffs between specialized agents to solve complex tasks. The system distinguishes itself as a language porting validator, using deterministic mocks and specifications to verify feature parity between different language implementations of a codebase. It further extends agent capabilities by acting as a Model Context Protocol
Coordinates the lifecycle and communication of secondary agents to solve complex, multi-step tasks.
TaskingAI est un orchestrateur d'agents IA et une plateforme d'application utilisée pour construire, déployer et mettre à l'échelle des applications natives IA. Il fonctionne comme un backend multi-tenant en tant que service, fournissant l'infrastructure pour héberger et gérer des instances d'agents IA indépendants à travers plusieurs utilisateurs ou organisations sur une architecture partagée. La plateforme dispose d'un constructeur de workflow visuel et d'une console de gestion de projet, permettant aux utilisateurs de configurer la logique des agents et de tester les workflows de conversation via une interface graphique avant de les passer en environnement de production. Le système orchestre les grands modèles de langage en standardisant les interactions entre les fournisseurs cloud et locaux via une interface unifiée. Il prend en charge la génération augmentée par récupération (RAG) en intégrant des sources de données externes et des plugins de recherche dans les workflows des modèles. Les capacités supplémentaires incluent la gestion de session avec état pour suivre l'historique des conversations et une architecture basée sur des plugins pour étendre les outils des agents.
Orchestrates language models, retrieval systems, and tools to execute independent, goal-oriented tasks.
Viper is a command and control infrastructure manager and post-exploitation framework designed for adversary attack simulation and security assessment. It functions as an orchestrator for penetration testing, combining a system for managing compromised hosts across multiple operating systems with tools for security workflow automation. The platform is distinguished by its use of large language model agents to coordinate red team tasks, automate data processing, and provide intelligent decision support. It includes a network pivot visualizer that uses directional graphs to map relationships an
Orchestrates multi-step AI agent workflows to automate repetitive red team tasks and decision support.
Magic est un environnement de productivité tout-en-un et une plateforme d'agents conçue pour déployer, orchestrer et gérer des workflows multi-agents. Il fonctionne comme un système de coordination qui répartit les tâches complexes à des agents spécialisés, servant à la fois de moteur de workflow et de système de gestion des connaissances qui synthétise les informations provenant de PDF, de sites web et de bases de données en assets numériques structurés. La plateforme se distingue par une suite de contenu multimodale capable de générer des livrables professionnels, incluant des assets graphiques haute fidélité, des diagrammes techniques et des présentations. Elle intègre une couche de routage qui fait correspondre des objectifs spécifiques au modèle de langage le plus approprié et permet la création de nouvelles capacités via des interfaces conversationnelles et des intégrations d'écosystèmes de compétences externes. Les domaines de capacités étendus incluent la gestion des connaissances en entreprise, où les données internes sont converties en travailleurs IA réutilisables, et une gestion complète des coûts avec un suivi budgétaire granulaire aux niveaux des départements et des utilisateurs. Le système fournit également un environnement collaboratif pour le partage de projets en temps réel et un framework de sécurité incluant une exécution en conteneur sandbox et des workflows d'approbation humaine pour les opérations à haut risque. La stack complète, incluant les clusters et les services, peut être installée sur des environnements macOS ou Linux privés en utilisant des scripts de déploiement automatisés.
Uses a central controller to manage task distribution and coordinate parallel execution of specialized agents.
Ce projet est un framework agentique et un orchestrateur Llama Stack utilisé pour construire des applications IA autonomes. Il coordonne l'inférence de modèles et l'exécution d'outils pour décomposer des objectifs complexes en chaînes de raisonnement multi-étapes et en boucles d'inférence continues. Le framework intègre un système de garde-fous de sécurité dédié qui filtre les entrées et sorties du modèle via des modèles de sécurité pour appliquer des restrictions de contenu au niveau du système. Il inclut également une couche d'intégration d'outils qui mappe les requêtes de fonctions générées par le modèle vers des définitions d'exécution externes pour effectuer des actions au-delà de la simple génération de texte. Le système fournit des capacités de génération augmentée par récupération (RAG) en extrayant des embeddings de documents depuis des bases de données vectorielles pour injecter des connaissances externes dans les prompts du modèle. Ces services sont exposés via une couche API unifiée qui prend en charge les scripts programmatiques et les interfaces de chat graphiques.
Manages state and continuous inference loops to decompose complex goals into multi-step agent workflows.
ReAct is an agentic workflow template and prompting framework for large language models. It implements a logic pattern that integrates chain-of-thought reasoning with external tool execution to solve complex, multi-step tasks. The framework uses an interleaved reasoning and acting logic, forcing the model to document its internal thought process before executing an action. This cycle of planning and acting allows the system to interact with external APIs or databases and inject real-world data back into the model context to refine reasoning paths. The project covers autonomous task execution
Orchestrates the interaction between reasoning cycles and tool use to manage stateful agent workflows.
Langroid is a multi-agent orchestration framework and tool integration suite designed for building complex AI applications. It serves as a multi-modal integration layer that connects diverse local and remote language models with an agentic retrieval-augmented generation system. The project distinguishes itself through a collaborative message-exchange paradigm, allowing specialized agents to delegate tasks hierarchically and coordinate via structured communication. It features an advanced state management system for conversational AI, including the ability to rewind and prune conversation hist
Coordinates multiple agents through a managed loop that processes turns and maintains state until completion sequences are met.
This project provides a comprehensive framework for building, deploying, and orchestrating autonomous agents within a decentralized network. It serves as a collection of patterns and examples for developing intelligent software entities capable of performing complex tasks, making decisions, and interacting with other agents to achieve shared goals. The framework distinguishes itself through its focus on multi-agent orchestration and decentralized communication. It enables the coordination of specialized agent teams that collaborate on workflows through structured messaging protocols, allowing
Orchestrates stateful agent cycles and tool interactions to manage natural language task execution.