Augment Code by Harness Reviews & Product Details

What sets Augment Code apart from other AI assistants I've used is its true codebase awareness. Instead of just looking at the open file or nearby tabs, its Context Engine actually understands how our entire project connects. If I ask it to update a data model, it knows which services, validation functions, and downstream files need to change along with it. It feels like pairing with a senior engineer who already knows our architecture. Review collected by and hosted on G2.com.

The initial codebase indexing and context ingestion can take a noticeable amount of time, especially when onboarding massive repositories or multi-service monorepos. While it only happens upfront, waiting for the Context Engine to fully map all dependencies before getting optimal suggestions can feel like a bottleneck on day one. Review collected by and hosted on G2.com.

The main thing I like is how easy it is to get going. Setup didn't take long, and the UI is clean enough that I didn't have to hunt around to figure out where things are. I've been using it for building and refactoring in a fairly large codebase, and it's easy to work with day to day without a big learning curve. Review collected by and hosted on G2.com.

During my initial setup it failed to create the cloud-based environment, so I ended up setting everything up locally instead. It works fine that way, but it wasn't how I planned to start. I've only been on it for a week or two, so I can't say much yet about how it holds up over the long run. Review collected by and hosted on G2.com.

The way it polishes code generated from other platforms. I bring in rough output and Augment Code cleans it up meaningfully, it doesn't just lint, it actually understands the intent behind the code and refines it accordingly. For a hybrid workflow where you're generating on one tool and perfecting on another, it slots in really well. Review collected by and hosted on G2.com.

The UI isn't the most intuitive out of the box. It takes a bit of trial and error to figure out where things are, which isn't ideal when you just want to get straight to work. A short walkthrough video or a quick-start guide would make onboarding a lot smoother for new users. Review collected by and hosted on G2.com.

I use Augment Code with Zed ACP, and it’s been superb. The ACE context engine is outstanding, and its awareness is genuinely impressive. It understands the project structure and the codebase really well, which makes working across files feel smooth and natural. On top of that, the terminal integration is solid and gives me a lot of freedom in how I work. Review collected by and hosted on G2.com.

The pricing is a little steep. I don’t have to worry about it because my company provides it, but on a personal level I would prefer Claude Code. Also, customer support can be a bit unresponsive. Other than that, it’s a great product. Review collected by and hosted on G2.com.

What I like best about Augment Code is how well it understands the context of a large codebase instead of just generating isolated snippets. It provides relevant code suggestions, explains existing code clearly, and helps with refactoring, debugging, and implementing new features without constantly switching between files. The responses are fast, the suggestions are accurate, and it integrates smoothly into the development workflow, making it easier to stay focused and be more productive. Review collected by and hosted on G2.com.

One area where Augment Code could improve is the occasional generation of suggestions that require manual verification, especially for complex business logic or edge cases. While it generally understands the codebase well, there are times when additional context is needed to produce the most accurate solution. I would also like to see more customization options for suggestions and even deeper support for large, enterprise-scale projects. Review collected by and hosted on G2.com.

The codebase-wide context lookup is its best feature. Unlike other tools which work on a file-by-file basis, Augment understands the cross-repository dependencies across modules. The time it saves to write unit tests and refactor the code according to the exact imports in our project is significant, whereas VS Code inline completion is fast and unobtrusive. Review collected by and hosted on G2.com.

Repository indexing at start can sometimes be long with big legacy code bases. Also, although the context understanding is great in most cases, some multi-step prompts sometimes may result in leaving small helpers or debugging statements which need to be cleaned up before creating a pull request. Review collected by and hosted on G2.com.

It actually understands repo-wide context instead of just guessing local syntax. The autocomplete feels instant in VS Code, filling out parameter types, multi-file importz, and matching internal helper functions without locking up the editor. It runs super light in the background, so you don't get that annoying input lag or cursor freezing when typing at full speed. Review collected by and hosted on G2.com.

The initial repo index takes a bit to catch up when opening large projects for the first time. Also, if your codebase has old or deprecated code paths lying around in secondary directories, it occasionally surfaces those legacy function signature in its suggestions unless you explicitly clean them out. Review collected by and hosted on G2.com.

It really understands how my project is structured. Rather than proposing some generic code snippets, it checks your repository and gives you completions that fit your particular codebase and coding style. Another huge advantage of the editor plugin is its extreme lightness–no lags and cursor lookups while typing quickly, as it's the case with bulky AI plugins. Review collected by and hosted on G2.com.

When you open a project for the first time, the initial index scan takes a few minutes to settle. Also if you have legacy or deprecated files sitting in the same repo, it sometimes pulls order function patterns into its suggestions unless you clean those folders up. Review collected by and hosted on G2.com.

Its multi-repository context awareness is exceptional. Rather than limiting analysis to individual open files, Augment maps dependencies and relationships across the entire project structure. It dramatically cuts down the efforts needed for refactoring and test writing by using precise local important, inline the online completion in VS Code remain quick seamless. Review collected by and hosted on G2.com.

Initial codebase indexing can take a bit of time when opening massive legacy projects. Additionally, while deep context retention is generally solid, complex multi-step prompts occasionally leave behind leftover print statements or utility snippets that require a quick manual cleanup before opening a PR. Review collected by and hosted on G2.com.

Augment Code's repo-wide context is top-tier, Refactoring design tokens or updating shared UI components used to mean breaking code across multiple files, but it catches all those dependencies easily. It actually follows our design system patterns instead of giving generic snippets, saving tons of time on multi-file refactors. Easily a game-changer for shipping clean UI fast. Review collected by and hosted on G2.com.

On complex or niche custom framework,it can still occasionally hallucinate or require manual checking of suggestions. The UI could also use a quick walkthrough, as finding some advanced context features takes a bit of trial and error at first. Review collected by and hosted on G2.com.