MCPFast / Tools / Python library for multi-provider LLM coding agents

GitHubAgent★★★★☆

Python library for multi-provider LLM coding agents

A strongly typed Python library and CLI for multi-provider LLM coding agents, with Erlang roots.

View on GitHub

sagent: Multi-Provider LLM Coding Agents for Developers

sagent is a robust Python library and command-line interface (CLI) designed for developers building multi-provider Large Language Model (LLM) coding agents. Drawing inspiration from Erlang's concurrency and fault-tolerance principles, sagent provides a structured and type-safe environment for orchestrating complex AI coding workflows. This tool is ideal for developers who need to integrate with multiple LLM providers and manage agent interactions efficiently.

What sagent Does

sagent simplifies the development of LLM-powered coding agents by abstracting away the complexities of interacting with different LLM APIs. It allows you to define and manage agents that can leverage the strengths of various LLM providers, enabling more flexible and powerful coding assistance. The library provides a consistent interface for sending prompts, receiving responses, and handling agent state, regardless of the underlying LLM.

Key Features

Who sagent is For

sagent is intended for software engineers , AI developers , and ML engineers who are building or experimenting with LLM-based coding assistants, code generation tools, or complex AI workflows that require interaction with multiple LLM services. If you are looking for a structured, type-safe, and reliable way to manage multi-provider LLM agents in your Python projects, sagent is a valuable addition to your toolkit.