Agents
See all 20 →Open-source desktop environment with mobile companions for running Claude Code, Codex, OpenCode, and other CLI coding agents in parallel. Each agent works in an isolated Git worktree, with integrated terminals, diffs, a browser, and remote-workspace support.
Claude Code is Anthropic's CLI-native agentic coding tool, running directly in the terminal with deep filesystem and shell access. It reads the entire codebase context and can autonomously edit files, run tests, and commit code — making it a favourite among power users who prefer a terminal-first workflow. Claude Code is widely regarded as the most capable agent for complex, multi-file refactors.
Aider is the most popular open-source terminal-based coding agent, enabling AI-assisted pair programming directly in your CLI. It maps your repository with a tree-sitter code graph and surgically edits files while automatically creating git commits for every change. Aider supports virtually every major LLM via API and is the go-to choice for developers who want full control without a GUI.
Amazon Q Developer is AWS's AI coding assistant, deeply integrated with the AWS ecosystem for building, deploying, and operating cloud applications. It can explain and generate Infrastructure-as-Code, answer questions about AWS services, and perform automated code transformations across the AWS SDK. Q Developer is the natural choice for teams already running workloads on AWS who want AI assistance without leaving their cloud.
Augment Code is an AI software-development platform for large codebases. Its plans include the Cosmos software factory, CLI access, daemon and automation workflows, MCP and native tools, IDE integrations, and AI review for GitHub pull requests, backed by Augment's codebase-aware Context Engine.

Bolt.new is a browser-based full-stack agent that scaffolds, runs, and deploys complete web applications from a single prompt — no local setup required. Built on StackBlitz's WebContainers, it has become the leading "vibe coding" tool for rapid prototyping and has spawned an ecosystem of open-source forks. It is especially popular with non-developers and startup founders who want functional MVPs without writing code.

Cline is an open-source coding agent available as a VS Code extension, CLI, and early desktop app for macOS, Windows, and Linux. It can edit files, run terminal commands, and use a browser with human approval checkpoints. Enterprise adds a JetBrains extension plus centralized security and administration controls.
Codex CLI is OpenAI's open-source terminal-based coding agent, positioned as direct competition to Claude Code and Aider. Running in your local terminal, it can read your codebase, write and edit files, and execute shell commands with configurable approval modes. As OpenAI's answer to Anthropic's Claude Code, it's generated significant interest among developers already in the OpenAI ecosystem.

Continue was an open-source coding agent for VS Code, JetBrains, and CLI workflows. Cursor acquired Continue, and the standalone service is no longer active. Its open-source codebase remains available for reference and reuse.
Cursor is the most popular AI-powered IDE, built as a fork of VS Code with deep model integration across autocomplete, chat, and inline edits. It supports multiple frontier models and has become the reference point for all coding agent comparisons. Its CMD+K and Composer features enable both single-file edits and full multi-file agentic workflows.

Devin is Cognition's software-engineering agent platform across Devin Cloud, CLI, and Devin Desktop, the renamed Windsurf IDE. It supports local and cloud agent fleets, integrated planning and review, browser and terminal execution, and multiple model providers including Cognition's SWE-2.
GitHub Copilot is the largest-installed AI code completion and chat tool, natively integrated into VS Code, JetBrains, and GitHub.com. Backed by Microsoft and OpenAI, it offers both inline suggestions and an agentic workspace mode. It remains the benchmark that all other coding tools are measured against in enterprise settings.
Goose is a local, open-source general-purpose agent governed by the Agentic AI Foundation at the Linux Foundation. It offers desktop, CLI, and API interfaces for coding, research, automation, and data work, with MCP extensions, skills, recipes, subagents, and configurable model providers.
Lovable is a no-code AI app builder that generates full-stack applications from natural language descriptions, with a visual editor and one-click deployment to its hosting infrastructure. It competes directly with Bolt.new but emphasizes design quality and Supabase integration for persistent data. Lovable is particularly popular for SaaS MVPs and has a growing marketplace of templates.
Open-source AI code-review CLI combining deterministic review pipelines with an LLM agent to produce line-level comments. Includes multi-language rules for issues such as null-pointer errors, thread safety, XSS, and SQL injection, with OpenAI- and Anthropic-compatible model support.

Qodo (formerly CodiumAI) is an AI coding agent focused on code quality, testing, and review — automatically generating meaningful unit tests and suggesting code improvements with explanations of behavior. Its Merge feature integrates into GitHub/GitLab PRs to provide intelligent code review with context from the entire codebase. Qodo is the go-to tool for teams that want AI to enforce quality standards rather than just write code.
Replit Agent is an AI coding assistant embedded in Replit's cloud IDE, capable of building full applications from prompts with a built-in package manager, database, and hosting. It is optimized for beginners and non-developers who want to build web apps without leaving the browser. Replit's collaborative multiplayer mode makes it a popular choice for hackathons and educational settings.

Roo Code was an open-source VS Code coding-agent extension with specialized modes and MCP support. The official extension shut down on May 15, 2026, and its archived repository should not be confused with the separate Roomote product now presented at roocode.com.

Sweep's former GitHub issue-to-PR bot has been retired from its original positioning. The official repository now points to a JetBrains coding assistant, while sweep.dev and its documentation site are unavailable. A community comment reports a shutdown, but no official shutdown notice was found.

Tricentis acquired Tabnine on July 30, 2026 to integrate its Enterprise Context Engine into the Tricentis Agentic Quality Engineering Platform. The former Tabnine site now redirects to Tricentis, and standalone self-serve pricing is no longer publicly listed.
Suggested comparisons
Categories
Reviews mentioning “coding”
I have used Claude Code daily for more than 6 months, mainly for coding and development work, and also a lot for writing research papers. The underlying Claude models are, in my opinion, some of the best out there, which is why the agent performs so well on both fronts. The one area where it clearly struggles is data science work, where it is noticeably weaker than it is at general coding or writing. On the models, I split tasks based on what they need: Fable handles planning and review, Opus is my primary model for heavy work, and Sonnet takes on lighter tasks that don't need much reasoning. The biggest downside is the subscription. Pro's limits were not enough for daily use, so I moved to Max, but that plan is expensive for a solo developer, and the usage caps have gotten stingier over time without the price changing to match. Despite that, it is still the agent I reach for every day.
Review for Claude Code
