dynamiq-ai/dynamiq

Dynamiq is an orchestration framework for agentic AI and LLM applications

What it solves

Dynamiq is an orchestration framework designed to simplify the development of generative AI applications. It provides a structured way to build, connect, and manage complex workflows involving Large Language Models (LLMs), AI agents, and Retrieval-Augmented Generation (RAG) pipelines.

How it works

Dynamiq uses a node-based architecture where different components (LLMs, agents, tools, converters, and embedders) are treated as nodes in a graph. These nodes can be connected in parallel or sequential flows using a Workflow class, which handles dependencies and data passing between them.

Key capabilities include:

  • Agentic AI: Support for ReAct agents that can use external tools (like code interpreters) and multi-agent systems where a manager agent can delegate tasks to specialized sub-agents.
  • RAG Pipelines: Built-in nodes for document conversion (PDF), splitting, embedding, and vector database integration (e.g., Pinecone) for both indexing and retrieval.
  • State Management: A GraphOrchestrator for creating custom architectures with states, edges, and conditional routing to handle iterative processes like feedback loops.
  • Memory: A memory module to maintain conversation history across sessions.

Who it’s for

Developers building AI-powered applications who need a more structured orchestration layer to move beyond simple prompts and create complex, multi-step agentic workflows or RAG systems.

Highlights

  • Flexible Orchestration: Supports parallel, sequential, and graph-based execution flows.
  • Multi-Agent Support: Enables the creation of manager-worker agent hierarchies.
  • Integrated RAG: Provides a complete pipeline from PDF ingestion to vector search and answer generation.
  • Tool Integration: Easy integration with external tools like E2B Code Interpreter and ScaleSerp.
  • Custom Routing: Conditional edges in graph orchestrators allow for dynamic, logic-based application flow.