This AI Tool Maps Any Codebase Before You Touch It (Understand-Anything)
BBetter Stack
컴퓨터/소프트웨어창업/스타트업AI/미래기술
Transcript
00:00:00You join a new team, you clone the repo, and you open the code base.
00:00:04And there it is, a 200,000 line code base looking at you like a deer in the headlights.
00:00:09So you do what every dev does.
00:00:11You grep, we jump between files, and we probably ask Claude to explain this repo.
00:00:17What if you got a map, a real interactive map that teaches you the code base?
00:00:22Flows, architecture layers, change, impact, this is understand anything.
00:00:27It's already hit over 14,000 stars on GitHub in weeks, and a lot of devs have been talking
00:00:32about it.
00:00:33In the next minute, I'll show you how this works and how it's going to immensely speed
00:00:36up your understanding of your code base.
00:00:44Understand anything is an open source Claude code plugin.
00:00:47It can also fit into workflows with tools like Cursor, Copilot, and Gemini CLI.
00:00:53You point it at a repo in even a knowledge base, and it turns that project into a queryable,
00:00:58interactive knowledge graph.
00:01:00It does this with static analysis plus multi-agent LLM processing.
00:01:05But the important part is not the tech stack.
00:01:07The important part is the part it actually solves, because every dev has had this problem.
00:01:13You are new to some kind of legacy code base, legacy app.
00:01:16The docs are outdated.
00:01:18The engineer who knew everything left six months ago.
00:01:22And your AI coding agent just keeps guessing over and over again.
00:01:26That is where this tool becomes really helpful.
00:01:28So let's run it.
00:01:29If you enjoy coding tools to speed up your workflow, be sure to subscribe.
00:01:33We have videos coming out all the time.
00:01:35Now here's a medium sized project.
00:01:36It's big enough to be really annoying, but small enough that I can actually show you how
00:01:40cool this is.
00:01:42It's just a repo I clone from Google microservices, and it's small enough that we can actually
00:01:46test this without having to pretend anything.
00:01:49First, I'm going to install the plugin, right here in Claude, plugin install, understand
00:01:54anything.
00:01:55Once this generates, we need to reload the plugins, of course, and then we just have
00:01:58to run understand.
00:02:00Now it's going to scan the entire repo.
00:02:03It pulls out structure, relationships, key modules, and likely business concepts.
00:02:09Now we can run the dashboard command to launch everything.
00:02:11Now first, this took ages to run, like 30 minutes, and it burned a boatload of tokens.
00:02:16So having a good Claude plan is a must.
00:02:19I have Claude Max, and this used 25% of my rate, so it burns and it burns fast.
00:02:24But once it is done, we can open this dashboard, and this is the really cool part that really
00:02:29hits home.
00:02:30I can zoom out and see high level architecture.
00:02:32I can zoom in and access internal parts.
00:02:35I can click to see the code breakdown and how all this code is connected.
00:02:40I can even click in and view the actual code itself.
00:02:43Then I can search for something here like payments.
00:02:46Now normally I'd be jumping between through routes, services, models, handlers, just docs
00:02:52that aren't even useful anymore, and here the tools pull the pieces together.
00:02:56Now I can click guided tour, and it walks through the flow in order.
00:03:01The entry point, the validation, the logic, database, external APIs, air handle it.
00:03:07That's already really useful.
00:03:09This guided tour breaks everything down.
00:03:11We can go into it, access the different components.
00:03:14That's the difference between, "Hey, this is a really cool graph, and I would actually
00:03:18check this before touching production code."
00:03:20But now the obvious question here is, don't we already have tools like this?
00:03:25As devs, we don't really need any more pretty dashboards.
00:03:29We need less wasted time, and this tool goes after one of the biggest time sinks in software.
00:03:35Getting context, right?
00:03:37Where is all this connected?
00:03:38How is it useful?
00:03:39People are testing it on large projects, legacy Java monoliths, microservice repos, hundreds
00:03:45of files, and the reaction is kind of what I got here.
00:03:48This would have saved me my first two weeks in the job because it breaks everything down,
00:03:53it connects everything, and it shows us how it's intertwined.
00:03:56That's the first really big use case, onboarding.
00:03:58So instead of saying, "Read these 12 pages and ask around if anything's confusing," that's
00:04:03already confusing, we can now say, "Open the graph, take the tour, then we can ask better
00:04:08questions."
00:04:09Now, the second really good use case is AI agents because most AI coding tools are only
00:04:14as good as the context that we give them.
00:04:17If the agent sees three files, it's just going to guess.
00:04:20If it has a structured map of the system with domains, flows, dependencies, and actual explanations,
00:04:27it has a better chance of making the right change in the first place.
00:04:30Then of course there's refactoring, right?
00:04:32Before we touch the code, we can now ask, "What does this code depend on?
00:04:36What flow does it belong to?
00:04:38What might break if it moves?"
00:04:40That is how you avoid turning a one-line change into a major event.
00:04:44And this is the real reason devs are pretty excited about this project.
00:04:48Not because we care about diagrams, I like diagrams, right?
00:04:51They're cool, they're useful, I'm visual, but we hate being lost.
00:04:55Now, this is where we need to be careful because devs have seen code visualization tools before.
00:05:00IDE graphs, source graph style navigation, NX graphs, tree setter visualizers, and a lot
00:05:05of them have the same exact problem.
00:05:09What do they do?
00:05:10They show structure, but they don't explain the actual meaning.
00:05:13They tell you this file imports that file, great, but why?
00:05:18What flow is this a part of?
00:05:19Where does the request start?
00:05:20What breaks if I change it?
00:05:22That's the missing layer.
00:05:24Understand anything is trying to add that layer.
00:05:27Instead of only showing as a TypeScript file or something else, it tries to turn into something
00:05:31closer to an actual flow of how things work.
00:05:35That's the big things here.
00:05:37From files to meaning, from imports to system behavior, from here are the pieces, to here
00:05:42is how the machine works.
00:05:43Compared to many LLM or rag code tools, it is also more visual and more teachable.
00:05:48A lot of AI code tools are basically just search boxes.
00:05:52You ask a question, you get an answer.
00:05:54This gives you a full on breakdown, helping us understand where everything is going, what's
00:06:00intertwined, how it's connected, with allowing us to actually see the code.
00:06:04Now let's talk about what people actually like.
00:06:05The obvious win here was onboarding.
00:06:07If you're joining a large code base, this gives you a good starting point.
00:06:12Also the architecture layers, right?
00:06:13These are really built out.
00:06:15You can start at the system level and then drill down into modules and implementation
00:06:20details.
00:06:21Diff impact is another big one.
00:06:23Every experienced dev knows the fear of making a tiny change in a code base they don't actually
00:06:27understand yet.
00:06:29Per Claude Code users, the graph can also become better context, so instead of dumping
00:06:34random files into a prompt, you give the agent structured architecture knowledge.
00:06:39It's also free, MIT licensed, incremental, and designed to work across multiple dev environments.
00:06:45Now on the skeptical side, when a project blows up this quickly, we start to ask the question,
00:06:49is this useful or did it just win the GitHub algorithm?
00:06:53That's a good question to ask.
00:06:55There are also the LLM dependencies, that means token cost.
00:06:59This blew up on token costs.
00:07:02So it took a while, it took a lot of tokens.
00:07:05Make sure you have a good plan if you're going to use this.
00:07:08You still need good judgment.
00:07:09This just gives you that overview.
00:07:11It doesn't replace reading code.
00:07:13It just helps you understand it better where everything is going.
00:07:16If you enjoy coding tools and tips like this, be sure to subscribe to the better stack channel.
00:07:20We'll see you in another video.
Community Posts
No posts yet. Be the first to write about this video!
Write about this video