Yup, Claude Code Plan Mode is here: The Senior Engineer's Workflow — Transcript
Full transcript
- 0:00Let's open claw code.
- 0:03Shift tab. Shift tab. Bam. We're in plan
- 0:07mode. The debate is over. Claw Code is
- 0:10taking over the engineering world. The
- 0:12Claw Code team continues their red hot
- 0:16streak as the most cracked engineering
- 0:19team. How are they able to do this? The
- 0:21creator of Claw Code said it himself.
- 0:24Generally at Anthropic, we have this
- 0:25product principle of do the simple thing
- 0:27first. The team has stealth launched
- 0:30plan mode. It's simple and powerful. How
- 0:33does plan mode work and why does it
- 0:35exist? There are two powerful ways to
- 0:38use plan mode. We're going to break down
- 0:41right now. One of these techniques takes
- 0:42a little more time but pays you
- 0:44dividends. You want to make sure you're
- 0:46using the right method so you don't
- 0:48waste your valuable engineering time.
- 0:50Let's master claw code plan mode in 20
- 0:53minutes.
- 0:59No one knows clawed code better than
- 1:02claw code. Let's ask it to tell us about
- 1:05plan mode. Reflect on your system
- 1:08prompt. Hi, can you explain plan mode?
- 1:12So, with certain features, you can be
- 1:13sure the system prompt has information
- 1:16about it. Here, we're quite literally
- 1:17asking Claude code to tell us about its
- 1:20new plan mode. You can see incoming
- 1:21tokens streaming in. And we look at
- 1:23that. Here is the details on plan mode
- 1:26right out of the system prompt. Plan
- 1:28mode is a special operating mode where I
- 1:30research and analyze, create a plan,
- 1:32present the plan for approval, wait for
- 1:34confirmation. So, this is really
- 1:35important and this is interesting. I
- 1:37didn't know it had an exit plan mode
- 1:39tool. This might be a fresh feature that
- 1:42just rolled out. The cloud code team's
- 1:44engineering velocity is mind-blowing.
- 1:46I'm restricted from making any file
- 1:47edits, running commands, creating or
- 1:50deleting, or making commits. And here we
- 1:52get a nice breakdown of when this is
- 1:54useful. We're working with AI here, so
- 1:56we can just ask stupid zero judgment
- 1:58questions that might have already been
- 2:00explained in the previous prompt, right?
- 2:01Ask stupid questions, validate your
- 2:03assumptions, learn fast, and move even
- 2:06faster. Can plan mode create or modify
- 2:09any files? Let's say like plan MD, can
- 2:13you modify that? So, I'm just being
- 2:15super super clear with cloud code. I
- 2:17want it to repeat things back to me.
- 2:19After all, we are operating in plan
- 2:22mode, so this is what it's for. No, plan
- 2:24mode cannot create or modify any files,
- 2:25including plan.MD. So, this is really
- 2:28about building up the context window
- 2:30inside of cloud code. There's no
- 2:32actions. There's no mutations happening
- 2:35here, right? So, let's be totally clear
- 2:37here. So I'll say so the purpose of plan
- 2:40mode read files understand the codebase
- 2:43by loading
- 2:45your context window with important
- 2:47information
- 2:49before we start working then switch out
- 2:52and edit correct me if I'm wrong. I like
- 2:56to use this correct me if I'm wrong to
- 2:59decrease sickle fancy basically decrease
- 3:01the bias in the model and present a more
- 3:04neutral question format. These models
- 3:06love to tell you exactly what you want
- 3:08to hear because it's good for business
- 3:09for these AI labs. So, you have to be
- 3:11really clear. So, I like to use this.
- 3:12Correct me if I'm wrong. It's just a
- 3:14little bit more neutral. You can see
- 3:15cloud code validating exactly what we
- 3:17were saying. Yes, you're exactly right.
- 3:18That's a perfect understanding of plan
- 3:20mode's purpose. It's research and
- 3:22planning phase where I load, analyze,
- 3:24design, present, then after you approve,
- 3:27exit, execute, and then we run all the
- 3:29commands so on and so forth. Right? This
- 3:31is really interesting. Notice how claw
- 3:33code is converging on the senior
- 3:36engineers developer workflow. First, we
- 3:38understand the feature by gathering the
- 3:40right information. We build up the state
- 3:42space of the problem and the solution.
- 3:45And only then do we start building.
- 3:47Before plan mode, we'd have to do
- 3:49something like this, right? You have
- 3:50your long prompt and then we say
- 3:52something like this, right? Don't write,
- 3:53update, or delete anything. Just read
- 3:56and plan. So with two shift tabs, we can
- 3:59immediately just enable that prompt
- 4:02modification. Planning mode means you
- 4:04can focus on priming your agent with the
- 4:07right context before getting work done.
- 4:09Great. So we have this new plan mode.
- 4:11Let's put it to work.
- 4:15I'm going to open up cursor here and
- 4:17let's enhance our previous proof of
- 4:20concept codebase, the infinite agentic
- 4:22loop. In this codebase, you can have
- 4:24claw code generate infinite versions of
- 4:26a solution to your problem until it
- 4:29breaks. Let's go ahead open up a new
- 4:31terminal window here. Fire up clog.
- 4:33We'll make sure we're running the
- 4:35state-of-the-art model opus 4. I'm going
- 4:37to run a prime command to get our
- 4:39context window loaded with key
- 4:41information about this codebase. You can
- 4:43check out the prime command inside of
- 4:45the custom slash command. This is a
- 4:47really simple one. We're just running
- 4:48get ls files and we're having cloud code
- 4:51read a couple of files. So you can also
- 4:52see we have claw.md added to this
- 4:55directory. So claw code is now primed.
- 4:56I'll hit shift tab tab. And now we're in
- 5:00plan mode. In the beginning I mentioned
- 5:02there are two ways to use plan mode. You
- 5:04can plan then write code or you can plan
- 5:07write the plan to a spec file and then
- 5:10have cloud code execute on the spec that
- 5:12you created. Four out of five times. I
- 5:15recommend the second technique
- 5:17specifically for mid to largesized
- 5:19features which you should be aiming for.
- 5:22We have more than enough compute and
- 5:23intelligent models in great tools. So
- 5:26aim big, hand off more work to your
- 5:28agent. Do more in single swings to save
- 5:31time and understand how far you can
- 5:34really push these models and these
- 5:36tools. So right now our infinite agentic
- 5:38loop set up in this codebase. If we open
- 5:41up source or source infinite, you can
- 5:43see they're generating single files. And
- 5:45in our previous video, we looked at
- 5:46these UIs, we can quickly just take a
- 5:48look at one of these. If I open up UI
- 5:50hybrid one, copy the path to this, and I
- 5:54just type Chrome this inside of bash
- 5:57mode, which you can trigger inside of
- 5:59cloud code with the bang or the
- 6:01exclamation. You know, you can take a
- 6:02look at this. Basically, we're just
- 6:03generating infinite UIs that could do
- 6:05many different things. If I go to two,
- 6:08you can see here's another one. Here's
- 6:10another one. Right? So, we just
- 6:13generated all of these interesting user
- 6:15interfaces.
- 6:17Right? Here's another kind of wild one.
- 6:19Here's another. Right? We can go, you
- 6:22know, we had tons and tons of these
- 6:23generated. If if you open up these
- 6:24directories, you can see all of them.
- 6:26You can check these out yourself. Okay.
- 6:27So, we ran a prompt that generated a ton
- 6:29of UIs inside of a single HTML file.
- 6:32Okay. So, very cool. What we're going to
- 6:33do here is update our lower level prompt
- 6:36that we passed into our infinite agentic
- 6:39loop prompt and we're going to have
- 6:40cloud code plan out exactly what we want
- 6:43done here. So what we're doing here is
- 6:44effectively metaprompting. I'm writing a
- 6:46prompt that's going to write a prompt
- 6:48that is a plan aka a spec. What I want
- 6:51to do here is just something simple. I
- 6:53want to showcase the power of plan mode.
- 6:55And instead of writing single
- 6:57self-contained HTML files, what we'll do
- 7:00is we'll break them up into their own
- 7:01isolated directory with a CSS.js and
- 7:05HTML file. Okay? And then we're going to
- 7:07embed all that in a new plan. So let's
- 7:10go ahead and get to work there. And all
- 7:11we need to do is be concise and write a
- 7:14great prompt. So I'll say let's plan
- 7:17specs V3. And I'll say V4. So I'm going
- 7:19to update that. That's what we want to
- 7:22generate. We want to build on specs v3.
- 7:26I'm using clause autocomplete here with
- 7:27the at@ symbol, but instead of writing a
- 7:30singlehtml
- 7:32file split into CSS.js.html,
- 7:37we also want to store these three files
- 7:40unique directory. All right. And that's
- 7:43it. So in plan mode, Claude Code is now
- 7:46going to get to work filling up its own
- 7:48context window with the information it
- 7:50needs to plan this workout. Okay, so
- 7:53we're going to kick this off. I'm going
- 7:54to leave open the specs directory.
- 7:56Nothing is going to get created here,
- 7:58right? This is all going into Cloud
- 7:59Codes context window. It's not going to
- 8:01create or modify anything, right? Just
- 8:03like it told us here, just like we
- 8:05learned by prompting and asking
- 8:07questions against Cloud Code. So let's
- 8:09go ahead and take a look at this. It's
- 8:10reading context. It's figuring things
- 8:12out. It's understanding what it needs to
- 8:14do to build out this V4 spec, right? So,
- 8:18it has a clear understanding of
- 8:19requirements and now it's going to
- 8:20present a plan. And so, this is quite
- 8:22interesting, right? It's planning inside
- 8:24of its own context window. It's just
- 8:27looking at your codebase, gathering
- 8:29information, reading files, right? It's
- 8:31all read. There's no mutations, there's
- 8:34no create, there's no update, there's no
- 8:36deletes. It's a very safe mode. And very
- 8:39importantly, this mirrors the beginning
- 8:41of a senior engineer's process, right?
- 8:45We plan, work first, we architect, we
- 8:47think, we plan, and only then do we
- 8:50build. So let's see what we have here.
- 8:52It's walking through based on my
- 8:53analysis. Here's a comprehensive plan.
- 8:55So this looks decent. You can see it has
- 8:57this key changes from V3 to V4. Each
- 9:00iteration creates its own all CSS,
- 9:02alljs. Perfect. So this looks good. Now
- 9:05I want to ask a question here about the
- 9:07implementation. There's one key detail.
- 9:08I want to make sure every version that
- 9:11we're going to generate with the
- 9:12infinite agentic loop gets their own
- 9:14directory with those three files. So I'm
- 9:16just going to be super clear here and
- 9:18showcase you know how you can use plan
- 9:20mode. So we're basically talking with
- 9:22the cloud 4 opus model running on claw
- 9:25code to be clear. showcase what the
- 9:28output durest
- 9:32mark directory looks like after running
- 9:36three agents. And so output directory is
- 9:39going to be part of our arguments, part
- 9:42of our variables inside of our infinite
- 9:44agentic loop prompt. And so I'm
- 9:46referencing this variable here, which is
- 9:48where all the iterations will be saved.
- 9:50So I'm going to kick this off. I want to
- 9:51be super clear about what's going on.
- 9:53And importantly, this conversation we're
- 9:55having is getting added to the context
- 9:58window. Right? So, check this out. This
- 9:59is exactly what I wanted to see. Right?
- 10:01Scroll up. You can see output dur right
- 10:03there. However, that gets defined in the
- 10:06infinite prompt, right? As a variable.
- 10:08We then are getting our own separate
- 10:11versions. And this is exactly what I
- 10:13want to see. I'm just proving things out
- 10:15uh with cloud code, right? I'm planning
- 10:17with cloud code with powerful
- 10:19intelligence. all my abilities, all your
- 10:21abilities, they're now augmented by
- 10:24cloud code, by Agenta coding tools, and
- 10:26by these models. So, make sure you can
- 10:28tap into it, right? Planning mode really
- 10:31helps us tap into this. So, let's go
- 10:32ahead and kick this off. I'm going to
- 10:34get out of plan mode. So, I'm going to
- 10:35run shift tab, shift tab. I want to get
- 10:38into autoaccept yolo mode. And now we'll
- 10:41let cloud code rip build the v4 spec
- 10:44mirror v3. and be clear about the
- 10:47improvements
- 10:49we are making. And I'll throw in a think
- 10:51hard to trigger the reasoning model
- 10:54here. This is awesome. So, we're going
- 10:55to get our new V4 spec out of this.
- 10:58Okay. So, the important thing here is
- 11:00that we're having Cloud Code write a new
- 11:03plan. There's a big problem with if we
- 11:05just said, you know, build out write the
- 11:07code for this, right? We can't see the
- 11:09full plan without consistently prompting
- 11:11Claude code. And remember, agents are
- 11:13ephemeral. What does that mean? They're
- 11:15temporary. They're stateless. Their
- 11:17context window will be blown away. So,
- 11:19we can improve on plan mode by having it
- 11:22write out the full plan into the specs
- 11:24directory, one of the three essential
- 11:26directories for agentic coding. Okay.
- 11:29And then we can once this plan is
- 11:31written here, we can review it, we can
- 11:34iterate on it, we can tweak it, and we
- 11:36can commit it to our repository, which
- 11:38has many downstream positive side
- 11:40effects. We'll dig into future videos.
- 11:42Make sure you're subscribed so you don't
- 11:43miss out on actionable agentic coding
- 11:46ideas like this. As you can see, week
- 11:48after week, there's a lot of value we
- 11:51can create with powerful agentic coding
- 11:53tools like cloud code. So, we're going
- 11:54to get the spec here written. It's
- 11:56activated the reasoning model. We use
- 11:57the think hard information dense keyword
- 12:00encoded by the cloud code team. Where
- 12:02was that? Let me search for that. Up
- 12:03here. Yeah, right here. Right. I said
- 12:05think hard and this triggers the
- 12:08reasoning model. Right now, Claude 4
- 12:10Opus is running its thinking process,
- 12:13making sure that everything's super
- 12:15clear. You can see very importantly,
- 12:16we're mirroring again another IDK right
- 12:19there. We're mirroring the V4 structure,
- 12:21but adapt it for the multifile approach.
- 12:24This is key. And while it's loading
- 12:25here, we can open up, you know, UI V3.
- 12:28Again, definitely check out the previous
- 12:29video to really understand the infinite
- 12:31agentic prompt and what's really going
- 12:33on here. But, you know, you can see just
- 12:35at a high level what exactly is going
- 12:36on, right? We're generating uniquely
- 12:38themed UI components that combine
- 12:40multiple UIs into one elegant solution.
- 12:43So that's what we're doing there. And
- 12:44you can see we just got V4 created for
- 12:46us here. 360 lines. Let's go ahead and
- 12:48open it up and make sure it makes sense.
- 12:50Okay. So themed hybrid UI specification
- 12:53V4 looks awesome. Builds on V3 with a
- 12:56critical architectural improvement,
- 12:58separation of concerns, new modular file
- 13:00structure. Great. And so I just want to
- 13:02quickly look through this plan and I
- 13:04want to see this key detail right here.
- 13:06output requirements directory structure
- 13:08UI hybrid iteration number very
- 13:11important right this differentiates it
- 13:13from v3 where we just use the iteration
- 13:15number from the infinite agentic loop
- 13:17and then we just create thehtml here
- 13:19we're creating the directory fantastic
- 13:22looks great I'm catching cloud code here
- 13:24before it makes a mistake I'm going to
- 13:25hit escape uh this is one of the you
- 13:27know great reasons to see the thinking
- 13:29of your model it was about to create an
- 13:31updated infinite command which we don't
- 13:33need at all this is good I'm going to
- 13:34stop it right here then I'm going to
- 13:36boot up another agent here. Open up a
- 13:38new cloud code agent and we're just
- 13:40going to kick this off. Right? So, this
- 13:41plan looks good. It has all the details
- 13:44from V3, right? We just built on the V3
- 13:47spec. And so, now we're going to kick
- 13:48off this infinite agentic workflow with
- 13:50our new plan. Just to kind of hop back
- 13:52into that idea from last week, this is
- 13:55the power of having higher order
- 13:57prompts. What does that mean? It's just
- 13:59like higher order functions where in
- 14:01certain programming languages, you can
- 14:03pass a function into a function. We're
- 14:05doing that same thing here with auntic
- 14:07coding and with prompt engineer. We're
- 14:09passing a prompt into a prompt, right?
- 14:13The infinite prompt takes a prompt as a
- 14:15parameter that it can operate in. Okay,
- 14:18there's a ton of value here. We're just
- 14:20scratching the surface of what can be
- 14:21done with higher order prompts. More on
- 14:23that later. Again, make sure to
- 14:25subscribe so you don't miss that. Let's
- 14:26go ahead and kick this off. Right, so
- 14:27there's V3. Let's close that and let's
- 14:29kick off V4 in our new Cloud Code
- 14:32instance. Check the model. We always
- 14:34like that big beefy opus model. And
- 14:37let's go ahead and kick this off, right?
- 14:39So slash um infinite. This is a custom
- 14:42slash command inside of the directory.
- 14:43We'll hit tab. And now we need a couple
- 14:45of arguments. So we can go ahead and
- 14:47sync these up. We need the spec file v4.
- 14:52We need our output directory source
- 14:55group. And then we need count. How many
- 14:58iterations are we going to run? Uh let's
- 15:00just fire off 10. We're going to have 10
- 15:01new versions here created. Let's go
- 15:02ahead and fire this off. Now our
- 15:03infinite agentic loop is going to read
- 15:06the spec output to the source group
- 15:09based on whatever is inside of our spec
- 15:11and it's going to create 10 versions.
- 15:12Let's go ahead and let this rip. And
- 15:14we're doing all this thanks to planning,
- 15:17right? This is all thanks to plans. We
- 15:19have plans on plans, right?
- 15:20Custom/comands. Then we have our specs,
- 15:23right? Another prompt. We had, you know,
- 15:24cloud code in plan mode, right?
- 15:27Thinking, building up context. So much
- 15:30of what we do as engineers in the
- 15:32generative AI age in this kind of second
- 15:34phase is all about building up
- 15:37information sets and organizing
- 15:38information for our agents and really
- 15:41understanding what our agents can do.
- 15:43There we go. You can see we have two
- 15:45waves. We're going to run in two
- 15:46batches. Wave one is going to create
- 15:48five agents. Wave two is going to create
- 15:49five agents as well. You can see their
- 15:51themes. And now cloud code and opus mode
- 15:54is going to kick off five agents in
- 15:57parallel. Again, once you finish this
- 15:59video, check out the previous one to
- 16:01really understand the infinite agentic
- 16:04loop prompt. Here's the magical moment
- 16:06we've all been waiting for. We're going
- 16:08to get five parallel agents kicked off
- 16:11and they're going to start working for
- 16:13us. There we go.
- 16:18So, you can see we have that new
- 16:19directory and we also have those five
- 16:22new agents running in parallel here.
- 16:24Let's go and just open all these
- 16:26directories up preemptively and we're
- 16:28going to be able to watch all of these
- 16:30parallel cloud code agents work together
- 16:33to accomplish every version of this
- 16:36task. Okay? And so if somehow you've
- 16:38been watching this channel and you
- 16:40haven't been convinced of the importance
- 16:42of the fundamental principle of AI
- 16:45coding, great planning is great
- 16:47prompting. You should absolutely be
- 16:49convinced now. Okay. a leading AI lab
- 16:52ghost launched and used the tab key for
- 16:56the feature for quick access. It might
- 16:58seem like I'm micro analyzing, but this
- 17:00is what great engineering is, right?
- 17:02Every detail matters. Every product
- 17:04decision matters and every product
- 17:06decision says something. All right, so
- 17:08you can see five agents running in
- 17:10parallel doing their own isolated pieces
- 17:13of work based on our plans. um you know
- 17:16principal ad coding members and
- 17:17subscribers of the channel you've
- 17:19already known about planning and
- 17:20architect editor prompt chains and you
- 17:23know that key principle that the plan is
- 17:26the prompt right we've known this for
- 17:27some time now you already have this
- 17:29deployed into production you're already
- 17:31using this in cloud code every single
- 17:33day this is just a new you know mode of
- 17:36operating for us as engineers now right
- 17:38but if you're just hearing about this
- 17:39principle of a coding the plan is the
- 17:41prompt check out principled AI coding to
- 17:44learn the other Seven principles of AI
- 17:46coding link in the description. This is
- 17:48fantastic, right? This is working for us
- 17:50on our behalf. We have five agents here.
- 17:52Everyone's up to two tool uses. It's
- 17:54probably going to be three each and then
- 17:56we'll be able to dive in and actually
- 17:57look at what's going on here. Okay, so
- 17:59it looks like hybrid v2 is coming around
- 18:02the corner here. This is exactly what
- 18:04we've detailed in our plan. And thanks
- 18:06to Cloud Code in plan mode here, right,
- 18:09shift tab, shift tab, uh, we were able
- 18:11to very quickly load up its context
- 18:13window with a couple key details, right?
- 18:16Very concise prompts. You don't have to
- 18:18type a lot. You don't have to do a lot
- 18:20with these great tools, but you do need
- 18:22to know exactly what to say and what to
- 18:24ask. We focus on these key ideas of
- 18:26writing concise prompts on the channel
- 18:28all the time. All right, so it looks
- 18:30like all of our versions are getting
- 18:31filled out here. You can see task two
- 18:33complete, task three complete. Both
- 18:36agents had four tool uses each. They
- 18:38both consumed about, you know, 40k of
- 18:40Opus tokens. That means there's a lot of
- 18:43value in these tokens, right? Uh 3
- 18:45minutes each. Not too bad. All right.
- 18:47And all of our other tasks are now
- 18:48complete. These should all be, you know,
- 18:50ready for us to operate in. Now, our
- 18:52agent, you know, running about 400
- 18:54seconds here. 5.2K tokens on the top
- 18:56level agent. Um and now we're going to
- 18:58kick off wave two. So, I'm going to stop
- 19:00it here. I I want these tokens for some
- 19:03other engineering work. But you can see
- 19:04it's going to kick off wave two here.
- 19:06I'm going to go ahead and stop it here.
- 19:07And let's just go ahead and look at our
- 19:08results here. Right. So I can type um
- 19:11Chrome. Let's open up one of these
- 19:13files. Copy the reference with command
- 19:16shift R. And I'll just paste that here.
- 19:18This is going to open up our brand new
- 19:20index script style. Still with that key
- 19:23prompt of generating interesting new UI.
- 19:25So let's go ahead and kick this off and
- 19:27check this out. Right. Uh nature's
- 19:28answers. Uh plant your search seed. Very
- 19:31interesting. All right. So, we have this
- 19:32interesting search bar here. We have a
- 19:34filter. Interesting UI there. Let's go
- 19:35on to number two. And remember, these
- 19:37were all generated in parallel. So, this
- 19:39is another interesting one that's going
- 19:41to validate as we type. And phone
- 19:43number, credit card. You see we have
- 19:45complete progress at the bottom. This is
- 19:47a really clean minimalist UI. Looks
- 19:49great. Let's look at V3. Okay. And V4.
- 19:53Okay. Very cool. We have like a music
- 19:55player in this kind of retro neon style.
- 19:59Okay. So there's more versions here. You
- 20:01get the point. Um the prompt or should I
- 20:03say the prompts and this prompt system
- 20:06uh that we have with the infinite
- 20:07agentic loop. It's doing all this work
- 20:09for us. This is very powerful. This is
- 20:11all thanks to planning as a you know key
- 20:15engineering concept that's been embedded
- 20:17inside of Cloud Code, the best agentic
- 20:20coding tool in the game right now. This
- 20:22codebase is going to be available to you
- 20:24in the description just as last week.
- 20:26I'm going to add this new spec here. And
- 20:28I'll leave source group in here as well.
- 20:30Planning is a key element of software
- 20:32engineering. First, we think, then we
- 20:34plan, and then we build. All with our
- 20:38agentic coding tools. There's a tectonic
- 20:41shift happening in the engineering world
- 20:42right now. The winning engineers and
- 20:44teams are setting up their code bases
- 20:46and tools, not for themselves, but for
- 20:49their agents. You know where to find me
- 20:51every single Monday. Stay focused and
- 20:54keep building.
About this transcript
This page contains the full transcript of Yup, Claude Code Plan Mode is here: The Senior Engineer's Workflow by IndyDevDan, generated from the public captions YouTube serves with the video. The transcript has 3,945 words across 569 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.