I figured out the best way to vibe code
Matthew Berman showed 13 AI tools in this video on . 11 are verified available in the EU. Each row below jumps to the exact second. Part of week 2026-25.
Tools shown in this video, with timestamps
One of the creator's two primary coding agents; supports models from multiple vendors, was early with cloud agents, and offers automations, rules and worktrees. Cloud agents provide video and screenshots of changes. Cursor also announced a Git alternative built for agent-scale deployment.
Live demoThe creator's other primary coding agent, praised for its design and concise summaries of what it did. Supports automations, cloud agents and worktrees.
Live demoGreat but used less by the creator because he runs out of quota quickly. Uses CLAUDE.md instead of agents.md and has cloud agents.
Recommended coding agent harness alongside Devin and Cursor, with multi-model support and web interface for cloud environments.
Recommended as a fantastic option with its own harness and multi-model support.
Rules files that tell coding agents how to work, including workflow, commit style and personality. Supported by most tools except Claude Code, which uses CLAUDE.md.
Live demoReusable skills for agents covering repeated tasks, domain rules, tool instructions and quality gates. The creator showcases a public Agent Skills repo with 61,000 GitHub stars covering the dev cycle from PRD to deployment.
Live demoAI code review tool that automatically reviews PRs, giving a summary, a 0-5 confidence score, flowchart and fix prompts. Sponsor of the video.
Live demoFeature that triggers agents automatically, e.g. on a GitHub PR opened. The creator builds one that waits for Greptile comments, fixes them and pushes. Codex has a similar feature.
Live demoA free library of agent loops (trigger, repeated action, goal), announced in the video and hosted on here.now. Examples include overnight docs sweep, sub-50 ms page load loop and production error sweep.
Live demoFree instant web hosting for agents, used to host the loop library.
Cursor's own model, described as excellent at writing code from a plan, used in a multi-model workflow for execution.
Used in the creator's multi-model workflow as the code reviewer to give an alternative viewpoint.