Skip to main content

Introduction

Welcome to BindAI. BindAI is a modular Python framework for building intelligent AI applications with reusable agents, tools, workflows, memory, and retrieval capabilities. Instead of writing AI applications from scratch every time, BindAI provides a consistent architecture that lets you compose production-ready AI systems from small, reusable building blocks.

Why BindAI?

Modern AI applications quickly become difficult to maintain. A simple chatbot often grows into a system that needs:
  • multiple agents
  • external tools
  • long-term memory
  • document retrieval
  • workflow automation
  • scheduled tasks
  • human approvals
Without a clear architecture, these features become tightly coupled and difficult to extend. BindAI solves this problem by separating every responsibility into independent modules.

Core Principles

BindAI is built around several design principles.

Modular

Every component can be used independently. You can use only the Agent package, only the Workflow engine, or combine everything together.

Framework Agnostic

Although BindAI includes integrations for popular providers, the framework itself is provider-independent. The same agent can work with different AI providers by simply changing the configured model provider.

Production Ready

BindAI was designed with production applications in mind. Features include:
  • structured execution
  • reusable workflows
  • tool calling
  • event system
  • middleware
  • memory
  • retrieval
  • scheduling
  • retry policies
  • timeout handling

Python First

BindAI embraces modern Python development. Everything is built using Python classes and familiar object-oriented patterns. No custom scripting language is required.

Main Components

BindAI consists of several independent packages.

Agents

Agents communicate with language models and orchestrate AI execution. They can:
  • maintain conversations
  • call tools
  • access memory
  • retrieve knowledge
  • produce structured outputs

Tools

Tools allow agents to interact with external systems. Examples include:
  • calculators
  • APIs
  • databases
  • search engines
  • custom Python functions

Memory

Memory enables agents to remember previous interactions. Providers can be swapped without changing the agent itself.

Knowledge

Knowledge sources provide retrieval capabilities for documents and structured information.

Workflows

The workflow engine orchestrates complex execution logic. Supported patterns include:
  • sequential execution
  • conditions
  • loops
  • parallel branches
  • retries
  • scheduling
  • human approval
  • timeout handling

Projects

Projects organize complete AI solutions. A project may contain:
  • applications
  • agents
  • workflows
  • shared tools
  • configuration
  • knowledge
  • memory providers

Documentation Roadmap

If you’re new to BindAI, the recommended learning path is:
  1. Installation
  2. Quick Start
  3. First Agent
  4. Tools
  5. Memory
  6. Workflows
  7. Projects
Following this order introduces the framework progressively while keeping each concept easy to understand.