Agent V

Agent V is an AI coding agent tool that runs on your desktop and automates software development tasks inside your existing repositories. You describe what you want and define success criteria, and the agent plans the work, reads and edits files, runs terminal commands and tests, and manages git operations. Every task record shows which files changed, which commands ran, and how many tokens were used, and you review the changes before merging, with the ability to rewind to any checkpoint. It integrates with Playwright for browser-based UI testing, GitHub for pull requests and issues, and MCP servers such as Gmail and Google Drive, and it works with 11 or more AI providers including OpenAI, Anthropic, and local Ollama models using your own keys. Governance features include allow/ask/deny permission rules for commands and edits, per-task spending and token tracking, and sandboxed command execution with timeout limits. Agent V suits development teams that want to delegate repetitive coding work while keeping code review and quality control. For anyone comparing AI tools for coding automation, it emphasizes auditability and self-hosting with your own model keys.

Automating coding tasks in repositories: writing features, fixing bugs, refactoring, running tests, and opening pull requests with full change logs