Back to browse
Parley – code review TUI for AI code

Parley – code review TUI for AI code

by pi-victor·Jun 2, 2026·2 points·0 comments

AI Analysis

●●SolidNiche GemCozy

Terminal-first AI code review with MCP integration for agent feedback loops.

Strengths
  • Explicit thread states (open, pending, addressed) track AI review progress
  • MCP server lets external agents address review comments programmatically
  • Customizable prompts for different agent harnesses (codex, claude, opencode)
Weaknesses
  • Narrow audience: terminal users doing local AI code review specifically
  • Author admits still fixing couple of things, not fully polished yet
Target Audience

Developers using AI coding agents locally

Similar To

GitHub Reviews · GitLab MR · Claude Code

Post Description

Wrote my own tool that helps me review AI Slop. Akin to Github/GitLab reviews but local and in a TUI instead. You get full review history, branch, worktree, commit switching and supports AI agents in codex, claude, opencode and pi. You can also customize the prompts that get sent to these agent harnesses when calling the agent from the TUI.

It has an MCP as well, so you can just tell your AI agent to address the comments in review "new-review" for example, instead of running the AI agent inside of the TUI.

I've been daily driving this for a while now, while i still need to fix a couple of things, but otherwise it's my favorite tool.

Similar Projects

Developer Tools●●●Banger

Deff – Review AI-generated code changes

Purpose-built diff reviewer for AI-generated code, vim-native, actually ships.

Solve My ProblemWizardry
flamestro
103mo ago