dynamiq-ai/dynamiq

What it solves

Dynamiq is an orchestration framework designed to simplify the development of complex AI applications. It provides a structured way to build and manage retrieval-augmented generation (RAG) pipelines and agentic AI systems, reducing the manual effort required to connect LLM nodes, tools, and memory modules.

How it works

The framework uses a node-based architecture where different components (LLMs, agents, converters, splitters, and vector stores) are treated as nodes in a workflow. These nodes can be connected in parallel or sequential pipelines using a Workflow class, which handles dependencies and data flow. For more complex logic, it offers a GraphOrchestrator to create state-based architectures with conditional edges and routers for iterative processes.

Who it’s for

It is intended for developers building Gen AI applications who need a scalable way to orchestrate multiple agents, manage RAG indexing and retrieval flows, or create custom agentic workflows with memory and tool integration.

Highlights

• Multi-Agent Orchestration: Supports parallel, sequential, and hierarchical (manager-sub-agent) agent configurations.
• RAG Pipeline Tools: Integrated nodes for PDF conversion, document splitting, embedding generation, and Pinecone vector storage/retrieval.
• Flexible Orchestration: Offers both simple linear workflows and a complex Graph Orchestrator for state-machine style logic.
• Tool Integration: Built-in support for tools like E2B Code Interpreter, ScaleSerp, and AWS Bedrock AgentCore.
• Conversation Memory: Includes a memory module with backend support (e.g., in-memory) to maintain chat history.