Best alternatives to Pawse

Build an AI Agent (From Scratch)

#1

Practical guide to designing and coding autonomous AI agents

This Manning Publications book provides a step-by-step approach to building autonomous AI agents that can plan, use tools, browse the web, and execute multi-step tasks. It covers agentic frameworks, memory systems, tool calling, and evaluation strategies with hands-on Python implementations. The book is aimed at developers and AI engineers who want to move from passive LLM usage to building production-ready agentic systems.

Claude Code

#2

Anthropic's agentic AI coding tool that lives in your terminal

Claude Code is an official agentic CLI tool by Anthropic that brings Claude directly into the developer's terminal and editor. It can read, write, and refactor files across a codebase, run shell commands, and execute multi-step coding tasks autonomously. Built for developers who want a powerful, safe AI coding agent without leaving their existing workflow.

Cursor

#3

AI-first code editor built on VS Code with deep model integration

Cursor is a fork of VS Code that embeds powerful AI features including multi-file editing, codebase-wide chat, and an AI agent that can autonomously complete tasks. It supports GPT-4 and Claude models, autocomplete, and inline diff review. Popular among professional developers who want AI assistance tightly integrated into a familiar editor.

Devin

#4

The first fully autonomous AI software engineer for real tasks

Devin by Cognition AI is an autonomous AI software engineer that can plan, code, debug, and deploy full projects end-to-end inside its own sandboxed environment with a browser, terminal, and editor. It tackles multi-step engineering tasks from a single natural language specification. Designed for engineering teams who want to delegate entire coding tasks to an AI agent.

Gemini CLI

#5

Google's open-source AI agent for coding tasks in the terminal

Gemini CLI is an open-source command-line AI agent by Google that uses Gemini models to assist with coding, file editing, and agentic tasks directly in the terminal. It connects to Google Search for grounding and supports tool use for real-world actions. Aimed at developers who want a capable, free, open-source terminal AI assistant.

google.github.io/adk-docs/a2a/

#6

Google's Agent-to-Agent protocol for interoperable AI agent networks

The Agent-to-Agent (A2A) protocol is Google's open specification, part of the Agent Development Kit (ADK), that defines how AI agents discover, communicate, and delegate tasks to one another across different frameworks and vendors. It complements the Model Context Protocol by focusing on agent interoperability rather than tool connectivity, enabling complex multi-agent orchestration. The protocol is aimed at developers building enterprise-grade agentic systems that span multiple specialized agents.

Hermes Agent

#7

Nous Research's open-weight agentic model for tool-use tasks

Hermes Agent is an agentic AI system by Nous Research built on their Hermes series of open-weight models, optimized for function calling, tool use, and structured outputs. It is designed to handle complex multi-step tasks with reliable JSON output and API interaction. Suited for researchers and developers who want a capable, open-weight model for agentic pipelines.

OpenCode

#8

Terminal-based AI coding agent powered by any LLM provider

OpenCode is an open-source terminal AI coding assistant that supports multiple LLM backends including OpenAI, Anthropic, and local models. It provides an interactive TUI interface for chatting about code, running edits, and executing terminal commands within a project context. Suited for developers who want a lightweight, provider-agnostic agentic coding CLI.

Openwork

#9

AI agent for automating knowledge work tasks across your tools

Openwork is an AI agent platform that automates repetitive knowledge work by connecting to workplace tools like email, calendars, and documents to complete tasks on behalf of users. It uses agentic reasoning to handle multi-step workflows end-to-end. Targeted at professionals and teams looking to delegate routine digital work to an AI assistant.