Atomic Agents

Active

Description

Atomic Agents is a modular AI agent building framework with an atomic design philosophy, providing composable components including tools, pipelines, and memory management for constructing agent systems.

Key Features

Atomic design philosophy with single-purpose, reusable, composable components
Built on Instructor and Pydantic for type-safe, predictable agent behavior
Multi-provider support including OpenAI, Anthropic, Groq, Gemini, and more
Dynamic context providers for runtime system prompt injection
Agent chaining with typed input/output schemas for complex pipelines
CLI tool 'Atomic Assembler' for downloading and managing tools and agents

Use Cases

💡 Building reliable, production-grade AI pipelines with predictable outputs

💡 Creating modular chatbots with custom tools and context providers

💡 Developing multi-step AI workflows with typed agent chaining

💡 Rapidly prototyping AI applications using composable building blocks

Strengths & Limitations

✅ Strengths

• Actively maintained, recent updates
• High community interest (6.2k stars)
• Permissive open-source license (MIT)
• Established track record (2 years in production)

Quick Start

Install: `pip install atomic-agents`. Install provider: `pip install instructor[anthropic]`. Define input/output schemas with Pydantic, configure SystemPromptGenerator, create an AtomicAgent, and call `agent.run(input_schema)`.

Visit GitHub

Related Projects

AutoChain

1.9k · Python

Stale C

AutoChain is a lightweight, extensible, and testable LLM Agent framework by Forethought, providing clean abstractions for agent building with automatic tool selection, conversation history management, and automated testing workflows.

frameworkagentpython +3

· Lightweight and extensible generative agent pipeline with minimal abstraction layers
· Support for custom tools and OpenAI function calling with automatic spec conversion
· Simple memory tracking for conversation history and tool outputs with BufferMemory

View on GitHub

LLMCompiler

1.9k · Python

Stale B

LLM Compiler for Parallel Function Calling (ICML 2024), significantly improving agent tool calling efficiency and speed through parallel execution.

frameworkpythonllm +2

· Automatically decomposes tasks into parallel and interdependent function calls
· Supports both open-source (LLaMA, vLLM) and closed-source (GPT) models
· Streaming mode reduces latency by dispatching tasks before full plan generation

View on GitHub

AIOS

6.3k · Python

Normal B

AIOS: AI Agent Operating System - a foundational runtime for large-scale deployment and management of LLM agents with scheduling, memory management, and tool registration.

pythonagentframework +2

· AIOS Kernel — LLM-embedded OS kernel managing agent scheduling, context switching, memory, storage, and tools
· AIOS SDK (Cerebrum) — platform for agent development, deployment, distribution, and discovery
· Multi-framework support — compatible with ReAct, Reflexion, AutoGen, Open Interpreter, MetaGPT and more

View full details → View on GitHub

LlamaFactory

74.5k · Python

Active A+

Unified Efficient Fine-Tuning of 100+ LLMs & VLMs. Supports LoRA, QLoRA, RLHF and more for building custom agent models.

pythonllmagent +2

· 100+ model support - covers LLaMA, Qwen3, DeepSeek, Gemma, GLM, Mistral, Phi and over 100 large language models
· Full training pipeline - supports pre-training, SFT, RLHF, DPO, KTO, ORPO, PPO and complete training workflows
· Efficient fine-tuning - supports 16-bit full tuning, LoRA, QLoRA (2-8 bit) plus GaLore, BAdam, APOLLO and other advanced algorithms