Tag

TradingAgents is an open-source multi-agent trading framework by TauricResearch that uses LLM-powered specialized agents (fundamentals, sentiment, news, technical, researchers, trader, risk/portfolio manager) to collaboratively analyze markets and produce trading decisions. It is research-focused, configurable (LLMs, data vendors, debate rounds), and integrates with OpenAI, Alpha Vantage and yfinance for data.

Sim is an open-source platform for visually building, running, and deploying AI agent workflows. It offers a canvas-based flow editor, Copilot-assisted node generation, vector database integration for RAG, and self-hosting options (Docker, Ollama, vLLM) for local or cloud deployment.

UI-TARS Desktop is a native desktop GUI agent by ByteDance that enables multimodal, vision-language-driven control of local and remote computers and browsers. It provides precise mouse/keyboard control, screenshot-based visual recognition, cross-platform support, and integration with the Agent TARS ecosystem and MCP tools. It focuses on private/local processing and building human-like task completion workflows.

Verifiers is an open-source library from Prime Intellect providing modular components to build, evaluate, and train reinforcement-learning environments for LLM agents. It includes SingleTurn/MultiTurn envs, ToolEnv for tool-enabled agents, rubric-based reward functions, parsers, and integrations with prime-rl and common inference stacks for both small-scale evaluation and large-scale RL training.

OWL is an open-source multi-agent collaboration framework from CAMEL-AI for building an AI 'workforce' to automate complex real-world tasks. It includes rich toolkits (browser automation, document parsing, multimodal analysis, code execution), MCP (Model Context Protocol) integration, local web UI, and examples. OWL focuses on parallel execution, privacy-first design, and customizable multi-agent workflows, and achieved top open-source performance on the GAIA benchmark.

An open-source GitHub repository that collects system prompts, internal tool configurations, and model-related snippets for many AI products and developer tools (examples: Claude, Perplexity, Replit, VSCode Agent, NotionAI and more). It presents a large corpus (advertised 30,000+ lines) useful for prompt engineering, AI tooling research, and security awareness. The repo includes sponsorship/support info, donation addresses, and a security notice for AI startups.

NVIDIA NeMo Agent Toolkit is an open-source library for building, connecting, and optimizing teams of AI agents and agentic workflows. Framework-agnostic and extensible, it provides reusable tools, automatic hyperparameter tuning, profiling, observability integrations, evaluation tooling, a UI chat interface, and Model Context Protocol (MCP) client/server support for integrating tools across services.

ART (Agent Reinforcement Trainer) is an open-source reinforcement learning framework for training multi-step, LLM-driven agents using GRPO. It supports models such as Qwen2.5, Qwen3 and Llama, offers a serverless W&B Training integration to handle infrastructure, includes example notebooks and benchmarks, and provides a client/server training loop that saves LoRA checkpoints and enables instant inference deployment.

MLX LM is a Python package to run, generate with, and fine-tune large language models on Apple Silicon using MLX. It integrates with the Hugging Face Hub, supports quantization and uploading of models, low-rank and full-model fine-tuning (including for quantized models), distributed inference and training, streaming generation, sampling/custom logits processors, prompt caching, and a convenient CLI and Python API.

12-Factor Agents is an open-source guide and GitHub project from HumanLayer that lays out 12 engineering principles for building production-ready LLM-powered applications and agents. It covers practical topics such as prompt ownership, context-window management, structured tool outputs, execution state, control flow, error compaction, small focused agents, and stateless design. The repo includes in-depth writeups for each factor, examples, and links to talks and related resources. Code is Apache 2.0 and content CC BY-SA 4.0.

PageIndex is an open-source, reasoning-based document indexing and retrieval framework by VectifyAI. It builds hierarchical tree indices (table-of-contents style) for long documents and uses LLM-driven tree search to perform relevance-first retrieval—no vector DB and no artificial chunking. Designed for professional long-document QA and analysis, PageIndex emphasizes explainability, traceability, and human-like navigation.

DroidRun is an open-source framework for controlling Android and iOS devices through LLM-driven agents. It enables automation of device interactions via natural language commands, supports multiple LLM providers (OpenAI, Anthropic, Gemini, Ollama, DeepSeek), offers planning for multi-step tasks, screenshot-based visual understanding, execution tracing, and an extensible Python API for custom automations.