Diffsmith

Comment on your AI agent's code & collaborate on changes

Developer Tools

Artificial Intelligence

Vibe coding

Visit WebsiteSee on Product HuntApp Store ⧉

the archiving-on-rewrite answer above covers the case where the agent edits a commented line. curious about the other trigger though: if the agent just runs `git commit` mid-session, does Diffsmith diff against the new HEAD and archive everything the same way, or does it lose track since it was anchored to the previously-uncommitted state?

finally a way to actually talk back to my AI agent line by line instead of copy pasting diff hunks into chat. The MCP reply viewer at the exact line of code is genuinely useful.

The MCP path is what makes this a loop rather than a one-way review, since the agent's replies land back at the line they belong to. Where does that comment state actually live: inside the repo as something I would gitignore, or in app-local storage outside the working tree? I ask because I want to know what happens to anchored comments when I switch branches or stash mid-review, and whether the MCP server is something the agent polls for new notes or you push to it when I finish a pass.

One thing I'd love to see is a keyboard shortcut to jump between comments during review, something like cmd+] and cmd+[ to cycle through them. Right now clicking around to find each note breaks my flow when I'm going through a big diff.

Finally a clean way to review what my AI agent just churned out without scrolling through terminal output. Clicking a line to leave a note and bouncing it back through MCP feels exactly like the workflow I didn't know I needed.

Would love to see a way to filter or group comments by the agent that generated the diff. When I'm reviewing changes from both Claude Code and Cursor in the same session, it gets hard to track which feedback needs to go back to which tool. A simple dropdown to view only comments for a specific agent's changes would save me a lot of mental overhead.

A diff view that groups changes by file with a quick-jump sidebar would save a lot of scrolling on larger branches, especially when I'm bouncing between several modified files at once.

Would love to see a way to group comments into review rounds so I can do a first pass, send notes to the agent, then do a follow-up review on just the changed lines from the next commit without losing my original feedback.

A diff view that groups related changed lines into collapsible logical blocks (function, class, block) would be huge when reviewing large AI-generated changes, instead of scrolling through every single line individually. Would love to see something like that next.

honestly this looks really useful for reviewing ai-generated code, the per-line commenting with mcp replies is a clever touch. one thing that would help me a lot is a way to filter the diff to only show files changed by the agent versus my own edits, so i can focus the review where it matters most. sort of a split view toggle.

The line-anchor model nails what the agent wrote, but reviewing Claude Code and Codex diffs my highest-leverage notes are almost always about what it didn't write: 'you skipped the error path here', 'no test for the empty input'. There's no changed line to click on for those. How do you attach a comment to an omission, or to a file the agent should have touched but left alone? That's where a lot of my review time actually goes.

Love how cleanly it anchors comments right to the diff line, makes giving feedback to the agent way less painful. The offline approach is a nice touch too.

One thing I'd love to see is a way to group related comments into threads so I can have a back and forth on the same block of code without the chat getting messy.

This is awesome, much needed visibility upgrade for reviewing what your coding assistant is ACTUALLY doing, love it.

honestly the line-anchored comments and pushing feedback back to the agent sounds super useful, but it would be even better if i could batch related comments into a single thread so i don't have to send five separate messages to the agent when i'm reviewing one function.

finally something that doesn't nuke my local diffs when I'm reviewing what cursor spit out, the line-anchored comments are kind of genius honestly

honestly this looks really useful for my workflow since i use cursor and claude code daily. one thing i'd love is a way to batch similar comments together, like if i want to flag the same issue across multiple files, it gets tedious clicking line by line for repetitive notes. a "apply comment template to selection" or something similar would save a lot of time when reviewing bigger changes

the line-anchored comments thing is genuinely clever, like you can basically hold a real conversation with the agent right on the code instead of copy-pasting snippets back and forth. also appreciate that it's fully offline, no telemetry nonsense, just open the repo and go

AI code review is most useful when it is tied to release risk, not just style feedback. I would rather see fewer comments with clear repro, affected behavior, and test evidence than a long list of plausible-sounding notes the team has to triage again.