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.