15+ AI tools for the terminal, from coding agents to local models

Want an AI model or a coding agent in your terminal? This list has coding agents such as Claude Code, Codex CLI, and Aider, tools that run models on your own machine such as Ollama and llama.cpp, and small helpers that send text to a model. Use the filters to narrow the list by app type, platform, language, or license.

showing 1–18 of 18 apps

appcategorystarspricinglicenseplatformslast commit
Runs an AI coding agent in the terminal with provider choice, sessions, tool use, and editor integration. ai and agents +1226.6Kusage basedMITlinux, macos, windowsNov 13, 2026
A local model runtime that runs large language models and exposes them over an API. ai and agents +1186.9KfreemiumMITlinux, macos, windowsNov 12, 2026
Anthropic's terminal coding agent that reads your repo, edits code, and runs tests. ai and agents +1155.7Ksubscriptionunknownlinux, macos, windowsNov 14, 2026
Runs and serves many language and multimodal model formats locally across CPU and GPU backends. ai and agents +1135.3KfreeMITlinux, macos, windowsNov 16, 2026
OpenAI's local coding agent that understands your repository and makes shell-level edits. ai and agents +1135.2KfreemiumApache-2.0linux, macos, windowsNov 14, 2026
Google's terminal coding agent for repo analysis and code edits on Gemini Code Assist plans. ai and agents +1108.8KsubscriptionApache-2.0linux, macos, windowsNov 13, 2026
Autonomous coding agent for the terminal and editor that plans, edits, and verifies changes. ai and agents +171.6KfreeApache-2.0linux, macos, windowsNov 12, 2026
Runs a terminal coding agent tuned for low-cost models, with editor support. ai and agents +168.9Kusage basedApache-2.0linux, macos, windowsNov 15, 2026
A modern Rust terminal with built-in AI, collaboration, and block-based output. terminals and shells +166KfreemiumAGPL-3.0macos, linux, windowsNov 15, 2026
An extensible coding agent that plans changes, edits repositories, and runs tools from the terminal. ai and agents +156.5KfreeApache-2.0linux, macos, windowsNov 13, 2026
slowA terminal coding assistant that edits a Git working tree and works with hosted or local models. ai and agents +150.1Kusage basedApache-2.0linux, macos, windowsMay 22, 2026
Applies reusable prompt patterns to text, video transcripts, files, and pipeline input. ai and agents +144.6Kusage basedMITlinux, macos, windowsNov 13, 2026
staleA terminal coding agent for planning and applying large multi-file changes with model context. ai and agents +115.9Kusage basedMITlinux, macos, windowsOct 3, 2025
Runs prompts, stores conversations, and manages plugins for hosted and local language models. ai and agents +112.8Kusage basedApache-2.0linux, macos, windowsNov 11, 2026
slowGenerates shell commands, code, and chat responses from prompts or piped terminal input. ai and agents +112.3Kusage basedMITlinux, macos, windowsJul 2, 2026
slowProvides chat, shell assistance, roles, sessions, and tool calling across several model providers. ai and agents +110.6Kusage basedMIT OR Apache-2.0linux, macos, windowsFeb 23, 2026
archivedSends pipeline input to language models for explanation, transformation, and command-line workflows. ai and agents +14,520usage basedMITlinux, macos, windowsMar 9, 2026
A terminal agent that works with files, shell commands, Python, browsers, and local or hosted models. ai and agents +14,478usage basedMITlinux, macos, windowsNov 14, 2026

How do you choose an AI tool for the terminal?

Most of these send your text or code to a model, so start with where that model runs.

  • Pricing. Coding agents are often free to install but charge for use, through a model provider's API or a subscription. Check the Pricing filter, then the provider's plans.
  • Local or cloud. Tools that run models on your machine keep your data local, but they need plenty of RAM and ideally a GPU.
  • Permissions. Agents can edit files and run commands. Check how each one asks before acting, and start in a project that's under version control so you can undo changes.
  • Platform. Check the Platform filter, and read the install docs for Windows notes.

What should you keep in mind?

Read what each tool sends to the provider and what it stores. API keys live in config files or environment variables, so keep them out of your shell history and out of any dotfiles you publish.