Codiff v0.1.0: A Local Diff Viewer for LLM-Generated Code Reviews
Codiff v0.1.0 is a native desktop diff viewer designed for developers who frequently review code written by LLMs. It provides a fast, minimal interface for reviewing staged and unstaged Git changes before committing.
Key Features
- LLM Walkthrough Mode: Run
codiff -wto get an LLM-generated walkthrough of the diff. - Inline Comments: Add comments on changed lines and copy the full review as Markdown with diff context.
- File Filters & Search: Filter by file and search within diffs.
- Large Diff Performance: Built to handle large diffs quickly.
Installation
Download the macOS app from the GitHub releases page. After installing, use Codiff > Install Terminal Helper to enable the codiff command.
The tool was built in 16 minutes by pointing an LLM at diffs.com and trees.software, and is available under an open-source license.
📖 Read the full source: HN LLM Tools
👀 See Also
memora: Version-Controlled, Typed Memory for AI Agents – Git for AI Beliefs
memora is a CLI tool written in Rust that version-controls AI agent memory with typed, provenance-tracked, branchable, and mergeable capabilities.
Selfware: Rust-based local AI agent framework with PDVR architecture
Selfware is an open-source AI agent framework built in Rust for local inference, implementing a PDVR cognitive cycle with 54 built-in tools and designed for long-running tasks on consumer hardware.
LLM Matrix: Community-Voted Model Comparisons Built with Claude Code
A data scientist built llm-matrix.vercel.app to compare LLM scores across multiple dimensions simultaneously, with community votes shaping rankings. The site was developed entirely using Claude Code with two specific plugins.
soul.py adds persistent memory to local LLMs with simple file-based approach
soul.py is a Python library that adds persistent memory to any LLM using two markdown files for identity and conversation logging, working with Ollama, OpenAI, and Anthropic models without requiring databases or servers.