Autono, developed by Vortezwohl, is a Python-based autonomous agent framework built on the ReAct paradigm. It enables agents to perform complex tasks through adaptive decision-making and multi-agent collaboration. Autono connects to external Model Context Protocol (MCP) servers and tools, functioning as a client framework that consumes these services. Distributed as source code via its GitHub repository, it supports various LLMs and connection protocols for agent operations. This tool is designed for developers creating advanced autonomous systems.

What does Autono connect to?

Autono connects to external Model Context Protocol (MCP) servers and tools, acting as a client or a framework for building clients. It integrates with these services rather than exposing its own tools as an MCP server. The framework supports standard input/output (stdio) for server connections, as well as Server-Sent Events (SSE) and WebSocket for different MCP server types. Autono agents utilize large language models such as gpt-4o-mini, qwen-plus, and deepseek-v3 during their operations.

Core features and tools

Autono is an autonomous agent framework based on the ReAct paradigm, designed for adaptive decision-making and multi-agent collaboration. Agents built within Autono can access various tools, including a calculator and the ability to write_file. The framework supports different task complexities: one-step tasks, multi-step tasks without failure, and multi-step tasks requiring retries and error correction due to possible tool failures. It also enables agents to balance conservative versus exploratory behaviors through hyperparameter tuning.

Who is it for?

Autono targets developers focused on building sophisticated autonomous agents. It allows for extensive customization of agent behaviors and supports scenarios requiring complex decision-making and collaboration among multiple agents. Developers can fine-tune agent tendencies, from cautious to experimental, by adjusting specific parameters. This framework is suitable for those creating systems that need to interact with external MCP services and execute multi-stage workflows. Skip this if your project does not involve building autonomous agents that interact with external MCP services.

How to get started

No credentials are required for basic use, as the official documentation does not detail an authentication model. Autono supports stdio for connecting to MCP servers, along with Server-Sent Events (SSE) and WebSocket protocols.

Use cases

• One-step tasks completed with a single tool call.
• Multi-step tasks requiring multiple tool calls without failure.
• Multi-step tasks requiring multiple tool calls where tools may fail, necessitating retries and error correction.
• Multi-agent collaboration scenarios.

Autono provides a flexible framework for developers to construct autonomous agents capable of handling intricate tasks and collaborating effectively. Its design supports various LLMs and connection protocols, making it adaptable for different MCP environments. The ability to customize agent behaviors and manage task complexities, including error recovery, makes it a valuable tool.