indusagi 0.2.1.1

A modern, extensible indus-agents for building autonomous agents with LLMs

indus-agents

A modern, extensible framework for building LLM-powered agents with tools, memory, and multi-agent orchestration. Built with Python and a clean, modular architecture.

Features

• Clean architecture with clear separation of concerns
• Type safety via Pydantic models
• Tool calling loop with built-in dev tools
• Multi-provider support: OpenAI, Anthropic, Groq, Ollama
• Multi-agent orchestration (Agency + Orchestrator)
• Conversation memory and tool usage logging
• Rich CLI and optional TUI

Quick Start

Installation

Using uv (recommended)

pip install uv
uv venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
uv pip install indusagi

Using pip

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate
pip install indusagi

Development install

git clone https://github.com/varunisrani/indus-agents.git
cd indus-agents
uv pip install -e ".[dev]"

Basic usage

from indusagi import Agent, AgentConfig

agent = Agent(
    name="Helper",
    role="Helpful assistant",
    config=AgentConfig(model="gpt-5-mini", temperature=0.7),
)

response = agent.process("What is 2 + 2?")
print(response)

Using tools

from typing import ClassVar
from indusagi import Agent, AgentConfig
from indusagi.tools import BaseTool, registry


class AddTool(BaseTool):
    name: ClassVar[str] = "add"
    description: ClassVar[str] = "Add two numbers"
    a: int
    b: int

    def execute(self) -> str:
        return str(self.a + self.b)


registry.register(AddTool)

agent = Agent(
    name="Calculator",
    role="Math helper",
    config=AgentConfig(model="gpt-5-mini"),
)

response = agent.process_with_tools(
    "Add 12 and 30",
    tools=registry.schemas,
    tool_executor=registry,
)
print(response)

Configuration

Copy .env.example to .env and fill in your keys. Example:

OPENAI_API_KEY=your-openai-api-key-here
ANTHROPIC_API_KEY=your-anthropic-api-key-here
GROQ_API_KEY=your-groq-api-key-here
OLLAMA_API_KEY=your-ollama-api-key-here

LLM_PROVIDER=groq
GROQ_MODEL=moonshotai/kimi-k2-instruct-0905
OLLAMA_MODEL=glm-4.7

DEFAULT_MODEL=gpt-5-mini
DEFAULT_TEMPERATURE=1
MAX_TOKENS=6000

CLI Commands

indusagi run "Tell me a joke"
indusagi interactive
indusagi list-tools
indusagi test-connection
indusagi list-agents
indusagi create-agent --output ./agents
indusagi tui
indusagi agency-demo
indusagi version

Project Structure

indus-agents/
  src/indusagi/
    agent.py
    agency.py
    orchestrator.py
    memory.py
    tools.py
    tools/
    providers/
    presets/
    templates/
    tui/
    utils/
    cli.py
  example_agents/
    example_agency*.py
    example_agency_improved_*_prompts/
  example-html-website-create-by-agents/
    <static html demos>
  useful-resource/
    Agency-Code/
    Mini-Agent/
    claude-agent-sdk/
  .env.example
  agents.md
  pyproject.toml
  README.md
  CHANGELOG.md
  CONTRIBUTING.md
  LICENSE

Examples and Demos

• example_agents/ contains the demo scripts and prompt folders used by the agency examples.
• example-html-website-create-by-agents/ contains static HTML/CSS/JS demo sites generated by agents.

Archived or Reference Projects

• useful-resource/ contains archived or reference subprojects (not part of the core package).

Development

uv pip install -e ".[dev]"
pre-commit install
ruff check src
ruff format src
python -m build

Contributing

Please see CONTRIBUTING.md for guidelines.

Support

• Issues: https://github.com/varunisrani/indus-agents/issues
• Discussions: https://github.com/varunisrani/indus-agents/discussions
• Email: [e-mail]

License

This project is licensed under the MIT License. See LICENSE.

Acknowledgments

• Built with Typer for the CLI
• Styled with Rich for terminal output
• Validated with Pydantic models

Citation

@software{indusagi,
  title = {indus-agents},
  author = {IndusAGI Team},
  year = {2025},
  url = {https://github.com/varunisrani/indus-agents}
}