Aider Alternatives

Free and open-source alternatives to Aider

• Codex CLIopen source
• Gemini CLIopen source
• Clineopen source
• Continueopen source
• Crushfree
• Opencodeopen source
• Qwen Codeopen source
• Spec Kitopen source
• Gooseopen source
• Serenaopen source
• Swe Agentopen source

Aider alternatives in detail

1. Claude Code

   paidCLI

   Claude Code is Anthropic's official agentic coding tool for the terminal, IDE (VS Code, JetBrains), and web. It works as an agent rather than autocomplete: it reads your codebase, plans changes, edits files, runs commands, and iterates against test or build feedback. Extensible via MCP servers, slash commands, subagents, and hooks; requires a paid plan.

   Read: Aider vs Claude Code: Open-Source vs Anthropic's Agent (2026)

   Read: AI Coding Agents in 2026: The Open-Source & CLI Edition

   Read: The Best Claude Code Alternatives in 2026 (Free, Open-Source, and Paid)

2. Codex CLI

   open sourceCLI

   Codex CLI is OpenAI's open-source (Apache-2.0) terminal coding agent: it reads files, edits them on disk, and runs shell commands in an OS-level sandbox with network off by default and, in trusted folders, workspace-scoped writes. Sandbox modes and approval policies control what it can do and when it must ask; sign in with a ChatGPT plan or API key.

   Read: AI Coding Agents in 2026: The Open-Source & CLI Edition

   Read: The Best Claude Code Alternatives in 2026 (Free, Open-Source, and Paid)

3. Gemini CLI

   open sourceCLI

   Gemini CLI is Google's open-source (Apache-2.0) terminal AI agent for Gemini models, with a 1M-token context window, file and shell tools, Google Search grounding, MCP, and GEMINI.md files. Since June 18, 2026 it no longer serves free or Google AI Pro/Ultra users, who are pointed to Antigravity CLI; paid API keys and enterprise Code Assist still work.

   Read: AI Coding Agents in 2026: The Open-Source & CLI Edition

   Read: The Best Claude Code Alternatives in 2026 (Free, Open-Source, and Paid)

4. Cline

   open sourceExtension

   Cline is an open-source autonomous coding agent that runs as a VS Code extension, with a JetBrains plugin and terminal CLI as well. Describe a goal and it plans, edits files, and runs commands, showing every change as a diff you approve before it executes. Bring your own model: Anthropic, OpenAI, OpenRouter, Google, Bedrock, or local runtimes like Ollama.

   Read: AI Coding Agents in 2026: The Open-Source & CLI Edition

   Read: The Best Claude Code Alternatives in 2026 (Free, Open-Source, and Paid)

5. Continue

   open sourceExtension

   Continue was acquired by Cursor, and as of September 2026 continue.dev is a farewell page rather than a product site. The Apache-2.0 codebase remains public: an extension for VS Code and JetBrains that supplied chat, tab autocomplete, inline edits, and agent modes while you brought your own model, hosted or local via Ollama.

   Read: AI Coding Agents in 2026: The Open-Source & CLI Edition

6. Crush

   freeFSL-1.1-MITmacOS, Windows, Linux, AndroidCLI

   Crush is Charm's free terminal coding agent. It continues opencode-ai/opencode (archived; last push September 2025), not today's OpenCode. It runs nearly any model, including local ones, switches models mid-session and adds LSP context, MCP and Agent Skills. It is source-available under FSL-1.1-MIT, not OSI open source; releases turn MIT after two years.

   Read: The Best Claude Code Alternatives in 2026 (Free, Open-Source, and Paid)

7. Opencode

   open sourceMITmacOS, Windows, LinuxCLI

   OpenCode is the most-starred open-source coding agent (~206k GitHub stars by September 2026) — a terminal TUI from Anomaly (formerly SST) that works with 75+ model providers including local ones, loads language servers for code intelligence, runs parallel sessions, and shares them via links. MIT-licensed; bring your own keys or use the optional Zen gateway.

   Read: Aider vs Claude Code: Open-Source vs Anthropic's Agent (2026)

   Read: AI Coding Agents in 2026: The Open-Source & CLI Edition

   Read: The Best Claude Code Alternatives in 2026 (Free, Open-Source, and Paid)

8. Qwen Code

   open sourceApache-2.0macOS, Windows, LinuxCLI

   Qwen Code is the Qwen team's free, Apache-2.0 coding agent, originally forked from Gemini CLI. It speaks OpenAI, Anthropic, Gemini and Qwen APIs, runs local models via Ollama or vLLM, and ships subagents, hooks, MCP and plan mode. The free Qwen OAuth quota ended April 15, 2026, so bring an API key, an Alibaba Cloud plan or a local model.

   Read: The Best Claude Code Alternatives in 2026 (Free, Open-Source, and Paid)

9. Spec Kit

   open sourceMITmacOS, Windows, LinuxCLI

   Spec Kit (GitHub, MIT, ~111k stars since its September 2025 launch) productized spec-driven development: specify init scaffolds a project, then /speckit.constitution, .specify, .plan, .tasks, and .implement walk your coding agent through the pipeline — each phase emitting a markdown artifact that feeds the next. Works with 30+ agents including Claude Code, Copilot, and Cursor.

10. Goose

    open sourceAgent

    Goose is an open-source (Apache-2.0), general-purpose AI agent that runs entirely on your machine — it executes shell commands, edits files, runs tests, and chains multi-step tasks. MCP-native with 70+ documented extensions and model-agnostic across 15+ providers including local models, it ships as a Rust CLI and a desktop app for macOS, Linux, and Windows.

    Read: AI Coding Agents in 2026: The Open-Source & CLI Edition

    Read: The Best Claude Code Alternatives in 2026 (Free, Open-Source, and Paid)

11. Serena

    open sourceMITmacOS, Windows, LinuxMCP

    Serena (MIT, ~29k stars as of September 2026) is 'the IDE for your agent': an MCP server backed by language servers giving agents symbol-level tools — find symbol, find references, replace symbol body, rename — across 40+ languages. Precise, token-efficient edits at the symbol level instead of regex surgery, plus a project memory for cross-session knowledge.

12. Swe Agent

    open sourceMITmacOS, Linux, WindowsAgent

    SWE-agent is a free, MIT-licensed autonomous coding agent from the Princeton NLP group that lets a language model of your choice fix bugs in real GitHub repositories. It pioneered the agent-computer interface (ACI) and posts strong SWE-bench results; the simpler mini-swe-agent now supersedes it for most use.