ANTIGRAVITY LABJP
Articles/Editor View
Editor View/2026-07-12Intermediate

Widening one panel made my agent reviews faster

A field note on turning two quiet additions in Antigravity v2.2.1 — Conversation Width and broader syntax highlighting — into a faster way to read the diffs your agents produce. Small display tweaks, but they changed how quickly I could review.

Antigravity324Editor settingsCode reviewProductivity4

Late one night, three agents I had running in parallel finished their tasks and stacked long diffs into the conversation panel. I started reading from the top, but as my eyes followed lines that wrapped sideways, I kept losing my place. The generation had gotten faster; my own reading speed had not kept up. That realization is what pushed me to revisit my display settings.

Antigravity v2.2.1 added Conversation Width (the maximum width of the conversation panel) to Appearance, and extended syntax highlighting in markdown code and diff blocks to C++, Python, and Protobuf. Neither is a headline feature. But for someone who spends long stretches handing implementation to agents, these two quietly decide whether I can actually finish reading what comes back.

When a line wraps, the review breaks apart

Diffs from an agent tend to run longer than patches a human writes. Generated code carries descriptive variable names and type annotations, and it is not unusual for a single line to pass a hundred characters. With the panel left narrow, those lines wrap mid-way.

The wrapping looks trivial. In a diff review, though, it stalls judgment. When I am tracking which added line corresponds to which removed line, a single line splitting into two breaks the vertical rhythm of the + and - markers. More than once I misread a wrapped line as a separate change for a moment. Re-read. Misread again. That back-and-forth adds up, tens of seconds per task.

Widening Conversation Width makes the wrapping disappear. A long diff fits on one line, and the column of +/- runs straight down. That alone let me grasp a block of change as a single surface, in one breath.

The right width differs per screen

Wider is not simply better, though. On my setup the ideal value shifted with display size and the kind of work. Here is where I landed after a few days of trying.

Working sceneDisplayWidth approachGoal
Diff review27-inch externalWide (kill wrapping)Read long lines in one row
Reading and writing plans27-inch externalMediumAvoid lines so long the eye slips
Laptop only14-inch built-inNarrow to mediumSkip horizontal scroll, follow the gist vertically

Prose and diffs pulled in opposite directions. Text that is too wide packs more characters per line, and I lose the start of the next line — the same reason newspapers use columns. Diffs, by contrast, get easier to read the wider they are. I settled on switching by time of day: wide on nights focused on diffs, medium during the day when I am reading prose like translation drafts.

Syntax highlighting shows the meaning of a change first

The other change, extended highlighting for C++, Python, and Protobuf, helped from a different angle than width.

With color, I know what kind of line it is before I read what changed in it. A type declaration, a string literal, a comment. Protobuf schema diffs were especially easy to miss when field numbers and types shared the same color as surrounding text. Once highlighting arrived, my eyes started stopping naturally on added and changed field definitions.

As an indie developer building apps on my own, when I use agents to assist that work I never trust a generated change outright — I always read it myself. And that act of reading, I only realized after changing my settings, depended heavily on display quality like color and width. Before debating how clever the agent is, put yourself in a state where you can read its output correctly. Unglamorous, but it matters more the closer you get to unattended operation.

This thinking gets more important when several agents stack changes at once. How to keep parallel work traceable after the fact is something I covered separately in a design for making parallel agent changes traceable later. Display settings are the groundwork that supports that tracing from the human eye's side.

The order I would revisit settings in

If your agent output feels hard to read right now, here is the order I tried.

First, widen Conversation Width one step and open a diff you usually review. Check whether the wrapping vanishes and the +/- align vertically. Next, return to a prose-heavy conversation and see whether the width is now too much; step back one if so. Finally, confirm highlighting is active for the languages you work in, using a real code block. A few minutes of tuning, but it keeps paying off across every diff you read afterward.

I also wrote about Antigravity splitting the IDE and the agent into separate apps in what tripped me up first when the IDE and agent became separate apps in Antigravity 2.0. The more the conversation panel becomes the main workspace, the more the readability of that one surface governs a day's focus.

The reader's speed is part of development speed

The faster agents generate, the more the bottleneck moves to human review. That is exactly why tending to the reader's environment — width, color — turned out to be a shortcut disguised as a detour. I am still searching for my own best settings. If today's diff reads a little easier, tomorrow's review gets a little lighter. That is the thought I keep in mind as I nudge these small settings.

Thank you for reading. If a small display tweak trims even one step from your reviews, I would be glad.

Share

Thank You for Reading

Antigravity Lab is ad-free, supported entirely by members like you. We publish practical guides daily with implementation code, benchmarks, and production-ready patterns. If you've found it useful, we'd love to have you on board.

  • Copy-paste ready implementation code
  • New advanced guides published daily
  • $5/mo or $10 for lifetime access
View Membership →

If you found this article helpful, a small tip ($1.50) would mean a lot to us. Your support helps keep this site ad-free and covers server and hosting costs.

Related Articles

Editor View2026-07-11
I Only Want to See the Lines the Agent Added: A Changed-Lines Convention Linter
I asked an agent for a small fix, it touched an old file, and the linter threw 13 warnings with no way to tell which were the agent's. Scoping the linter to only the added lines dropped 13 to 3 — about 77% less noise. Here is how to pull added lines out of a diff and check only those, with working code.
Editor View2026-07-09
Antigravity DevContainer: A Complete Setup Guide for Reproducible AI Development
An end-to-end guide to running Antigravity inside a DevContainer — Docker setup, Ollama integration, secrets, volume persistence, team distribution, and CI parity.
Editor View2026-07-02
Handing Off IDE Work to the Chat Agent: Passing Context Through a File After the Two-App Split
Since Antigravity 2.0 split into an IDE and a chat-style agent app, context you build up in one does not carry into the other. Here is a file-based approach that makes a single file in your repo the source of truth for handoff, with a schema, a validation script, and a way to pin it with a Guide skill.
📚RECOMMENDED BOOKS
Build a Large Language Model (From Scratch)
Sebastian Raschka
LLM Dev
Prompt Engineering for LLMs
Berryman & Ziegler
Prompting
AI Engineering
Chip Huyen
AI Eng
* Contains affiliate links
See all →