Hire Team
Home/Products/ Framework
BUILD YOUR OWN AGENTS

IndusAGI
AI Agent Framework

An embeddable agent stack for connecting models, tools, context, and workflows inside your own application.

TypeScriptPythonRust
Choose Your EditionExplore the Coding Agent
THE BUILDING BLOCKS

Your application.
Your agent workflow.

01

Connect your models

Use provider integrations through a common interface, with streaming responses and structured tool calls.

02

Build an agent workflow

Compose an agent loop, define tools, and manage the context your application needs.

03

Manage context and sessions

Keep agent context manageable through compaction and saved session history. Capabilities differ by edition; use its documentation to select the components your application actually needs.

FRAMEWORK, IN PRACTICE

An agent stack inside your application.

The IndusAGI Framework is for developers building their own AI-powered application. Its building blocks cover model access, streaming, tool calling and an agent runtime. The TypeScript documentation also covers MCP connections, context compaction and branchable sessions. Choose an edition after reviewing its APIs and supported integrations.

Context management is different from a vector database. In the TypeScript edition, the memory package name does not mean that a built-in VectorStore or Embedder is available. Plan retrieval integrations explicitly, and check the Python or Rust documentation rather than assuming feature parity.

Your deployment and model-provider choices determine where data is processed. Running part of a workflow locally does not make remote model calls local or establish regulatory compliance. Review data flows, tool permissions and retention for your use case.

Read the framework architecture
THREE LANGUAGE EDITIONS

Choose your stack.

Start with the edition that fits your environment. Each has its own repository, installation path, and documentation.

FRAMEWORK

TypeScript

A modular SDK for JavaScript and TypeScript applications, with model access, an agent runtime, tools, context management, and terminal UI building blocks.

npm install indusagi
Documentation GitHubPackage
FRAMEWORK

Python

An async-first Python framework for multi-provider model calls, tool-using agents, memory, and retrieval workflows.

pip install indusagi
Documentation GitHubPackage
FRAMEWORK

Rust

A Rust library and CLI with provider abstractions, agent orchestration, tools, memory, and terminal components.

cargo add indusagi
Documentation GitHub

Features, supported providers, and setup requirements vary by edition. Follow its documentation and repository for current configuration and license terms. Model-provider usage may have separate costs.

BUILD WITH INDUSAGI LABS

Need it shaped around your business?

Our services team can help turn an idea into an AI workflow, website, or custom software. Tell us what you want to build.

Discuss Your Project