YT YouTube2Text

YouTube2Text → Transcripts → ISU Fall 2026 - DGD205 Week 3 - 08.10.2026

ISU Fall 2026 - DGD205 Week 3 - 08.10.2026 — Transcript

by Onur EREREN · 9,193 words · 1,395 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:01Good morning everyone. Welcome to DGD
0:04205 game development pipelines and this
0:07is week three. It's 8th of October 2026.
0:12Now just before we started we were
0:14chatting a little bit about what a game
0:16engine is and why should we use one now
0:20that AI is here and we can just say put
0:23this on the screen do this and it just
0:25does it for us. Now
0:28I will just summarize and I will get
0:29back to the class after that. Now the
0:32end the thing that we call an engine.
0:35See I don't like this one. When I'm
0:37writing I'm just looking down in the
0:38camera here. So you see me looking down.
0:40You see the top of my head. It's
0:42pointless but it's in front of me. So I
0:44will probably drop this workflow but for
0:46this week bear with me. Anyway,
0:49now a game engine is not just a graphics
0:52engine. There is a for example there's a
0:54graphics engine called Auggre for that
0:56specific purpose. It just draws graphics
0:58on the screen. But a game engine does
1:01many things for you besides graphics. It
1:04for example Unity I'm proficient in
1:06Unity so I can tell you in Unity. Unity
1:09handles audio. It has mixers. It has
1:11spatial audio which means depending on
1:14where in 3D games depending on where the
1:16sound source is in 3D space and where
1:19your camera is it can adjust the left
1:21and right speakers so you hear it from
1:24that side. If you make your own engine,
1:26you have to program for that, right?
1:29Unity has animation pipelines, not very
1:33it's a little bit outdated and it's a
1:35little bit clunky and there are other
1:37packages out there that solve that
1:38problem. So you can buy them, integrate
1:41them. So other people are creating
1:44assets for these engines as well and you
1:46can use them. The other thing is
1:48ecosystem. You don't have to code
1:50everything. You don't have to AI have AI
1:53code everything, right? there's UI
1:55system there and there's a set code
1:57base. So for example, a person who only
2:00codes for Unity in C may not be
2:04proficient with regular software
2:07development or headless software
2:09development with C because Unity has
2:13update method and maybe you don't know
2:16how to advance a tech without that in C.
2:19Unity does that for you. So it gives you
2:21a framework to work with. Now when you
2:25want to do make your own engine, your
2:28custom engine, what you have to do is
2:30plan for all of these. You can plan for
2:33these step by step as the need arises.
2:36But every time you do, you will have to
2:38go into your codebase and rip things
2:42apart to make things fit. Now mate asked
2:45the question regarding this. There is AI
2:47now. Do we need to use those engines
2:51over there? But every new feature you
2:54want is going to be something you will
2:57tell AI to do for you, right? And AI at
3:03these at this this year and last year
3:06and probably next year as well,
3:09AI has the thing called tunnel vision.
3:12So it you tell AI to create you a UI
3:15system and it if it if it feels
3:18necessary it rips through the rest of
3:21your code base to make things fit to
3:24build the UI system but other things
3:27break. They do it all the time. Okay?
3:32Because they only focus at that moment
3:35that AI agent only wants to do what you
3:39wanted it to do. other things are
3:41collateral. Now what you can do to
3:44prevent this if you set out if you don't
3:48say
3:49draw a square on the screen it will do
3:51whatever it needs to draw a square on
3:53the screen okay and then you say I want
3:56to use wd to be able to move this screen
4:00move this square sorry
4:03so what it will do is just to create a
4:05basic system and for example it will
4:09bind it to w because you said so It will
4:12not include the arrow keys. It will not
4:13make it rebindable. Those things are not
4:15in its mind at all. And then you will
4:18say for example, I want space to shoot
4:22something.
4:23So we are gradually building a shoot
4:26them up, right? It's like this. And then
4:28I will replace this with a sprite. So
4:31but
4:33the AI doesn't know your next step, what
4:36you have in mind. It doesn't know you
4:39want to replace this with a plain
4:41graphic later on. So, it just didn't put
4:44that sprite idea in place. It just put a
4:47square. Right now, this is a very minor
4:49change. No problem at all. But when you
4:51want to switch them later on, it might
4:54need it might need to rip through
4:58previous things it has written.
5:03Does it make sense?
5:08Does it? I hope it does. So, what you
5:11can do is to start structured. If you
5:15want that, you tell AI, I want to build
5:19an engine that does this, this, this,
5:22and this. For example, you can say, do
5:24you have any questions for me? Because
5:26it might think of something you did not
5:28know about. It will probably think of
5:31something you did not know about because
5:32like it or not any AI agent that is chat
5:36GPT cloud or Gemini right now knows more
5:39than you about writing a game engine
5:41right normally. So so but you can talk
5:46with the AI to I'm going to show you
5:49something if only I can find this. Where
5:53is this? I'm looking for the place to
5:54put the stylus.
5:56So
5:58for example, let me show you something I
6:00did. This is you guys. I will share my
6:03screen with you with sound
6:09and I want you to see me as well because
6:12so for example actually let me open VS
6:16code. We will go better from there. I
6:18don't know which one will open. So I
6:20will just text detective. Okay, that's
6:22all right. But this one uh this one
6:25isn't the thing I'm looking for. I'm
6:29looking for undermine the incremental
6:32I'm working on. Here it is.
6:35So for example for undermine
6:38actually let me show you my GitHub if it
6:40is here.
6:43See, you see this screen right on the
6:46left hand side
6:52you have. Wait a second. Where is it?
6:56Where is it?
6:59Source. Here it is.
7:01So, this game, let me show you what it
7:05is.
7:09Where does it run?
7:11Show me where you run.
7:17Okay, it's not running. Huh? It is
7:20already. Okay.
7:2373.
7:29Now,
7:33this is this is the game. All right. So
7:36you just uh build shafts and then push
7:40mining equipment into it and it starts
7:41digging for you. So it's a mining
7:43incremental game. All right. Now I will
7:45show you this.
7:48The game is actually this.
7:52So I first built I told cloud that I'm
7:57going to use phaser and I want to build
8:00a hex game engine inside phaser. So I
8:05don't
8:07if there is a demo scene. Let me try to
8:10find that
8:17demo scene. There is a demo scene but I
8:19don't know how I can. So I have
8:21forgotten how I can do that. So
8:25let's ask
8:27I want to start the demo scene for this
8:32game.
8:35which will
8:37display all the underlying engine
8:41capabilities
8:43like tiles, arrows, etc.
8:49Scan the codebase and tell me how I can
8:54display it because I forgot. I mean,
8:57it's been about one and a half months
8:59since I made this. So it's going to
9:03look through things now.
9:06So this is a brand new fresh agent. It's
9:09going to go through the codebase. So uh
9:12level name public levels with no
9:14extension.
9:18So scene demo. It wants this one. Okay.
9:23Let's go back here.
9:30scene demo. Here it is. So,
9:35for example, I made this one and I
9:39worked with phaser. I worked with cloud
9:41code and I said I want a hex engine and
9:46for reference, let me show you how I
9:49structured this, how I planned this. I
9:51mean this is a little bit out of our
9:52scope for the class but I think this is
9:54a much more valuable thing for you to
9:57learn so we will catch up on the other
9:59don't worry about that where is my steam
10:01where is my steam now there is this game
10:04I don't know if you heard of it who who
10:07among you likes playing strategy games
10:11turnbased
10:13strategy trying to
10:23There is a game called Into the Bridge.
10:26If you don't know about it, it is on
10:28discount. Okay, you all have to buy it
10:29or you cannot pass my class. Okay, I'm
10:32not serious or or am I? But let me show
10:36you the trailer because it is just that
10:39good. It's worth one minute of your
10:41time.
11:00Heat. Heat.
11:07Heat. Heat.
11:17[music]
11:38>> [music]
11:42>> So you will understand why it is so
11:44great when you play it if you are into
11:46strategy games otherwise it might bore
11:48you. But every
11:51every game screen is like a chessboard
11:54and every turn what the enemies will do
11:57is perfectly shown. It's called
12:00telegraphing. So when you show what an
12:02enemy will do, it's telegraphing. You
12:04see it a lot in games. And in this one,
12:07there is no randomness. What each enemy
12:10will do is completely
12:13open to you. So you can see it, right?
12:16And then you act accordingly. So each
12:19turn is like solving a puzzle.
12:22Anyway, I just love this game. And I
12:25said to myself, I want to build a
12:27prototype like this, but with hexagons
12:30because this is a little bit I want to
12:32do try something different. I want I
12:34don't want to make a clone of this. So,
12:35I want to introduce something new. And a
12:38hexport is something that is tried and
12:40tested in strategy games. So, I said to
12:43cloud, I want to build some this is our
12:45gold standard. This is our baseline.
12:48Okay. And I want to build something like
12:49this. And it gave me a spec list. You
12:52should we need this, this, this, and
12:54this. And then I said, okay, first let's
12:57build the grid. And for example, I think
13:01besides the demo, there is this
13:05hexagon. No, hexagon is this one. Okay.
13:08Anyway, I don't want to go too deep into
13:11that. But what you can do is for example
13:15you can draw paths from one point to
13:18another
13:19like this. And these paths can follow
13:22around obstacles.
13:24So or you can highlight cells like this.
13:29And you can have twins move objects.
13:34So this is just a press of a button. And
13:38you can have indicators like this from
13:41one hex to another to show
13:42directionality, right? And you can have
13:45small icons on the bottom right or any
13:48point of the hex to show any status
13:51icons. So I thought what would a game
13:54like that require? And I said I want a
13:58game engine. Technically this is not a
14:00game engine itself. It uses phase phaser
14:03game engine but it's a another framework
14:06on top of it. Right? So this can build
14:10any game that uses a hex like this. This
14:13has no troop movement. It has movement
14:17but it doesn't have an understanding of
14:19a unit, a character, a stat. It is a
14:22grid. It can create a grid. It can draw
14:25lines on cells because I want to show
14:28paths for example. And it can calculate
14:31the path from one point to another, the
14:34shortest path. It's a very simple
14:36algorithm. It can highlight cells. It
14:38can highlight tiles. And it can print
14:41sprites on these tiles. So this is a
14:44game engine I built. And then on top of
14:49that,
14:53I I realized I first want to do
14:55something simpler. I don't want to go
14:56into complex unit calculations and
14:59combat mechanics and balancing. So I
15:02said I'll just make a mining incremental
15:04game. So this uses the same engine. The
15:06underlying engine is the same. And you
15:09can see now when I just for example
15:11click on this I highlighted the tile.
15:14This is engines capability. So I just
15:17say when I'm coding this game I say to
15:20the AI when I click on a tile it should
15:23be highlighted. That's all I have to say
15:25because the engine I built beforehand
15:29already has the code for it and most
15:31importantly
15:33I have full documentation for that here
15:38because one of my command commands to
15:41cloud was whenever we add a feature you
15:44will add extensive documentation for
15:47that because I will not always work with
15:50the same agent. Sometimes maybe I will
15:53use this game engine with another agent
15:56or I will give it to somebody else, a
15:58human to work with, right? Documentation
16:01has to be there. And in the older times
16:04when we were coding by hand,
16:05documentation was that thing everybody
16:07dreaded.
16:09No coders, designers, nobody wanted to
16:11mess with the documentation. They just
16:13wanted to do the job, right? But
16:15documentation is just as important as
16:18doing the thing. Without that you have
16:20no memory of what you did and why you
16:23did it. Imagine people just living
16:25through life old up till days and they
16:28never kept any records. It would be a
16:30mess, right? Our entire society would be
16:33a complete mess anyway. But now AI can
16:36do this very cheaply and I can highlight
16:39cells. This costs me nothing to
16:42implement in the game because the engine
16:44already has it. And for example, the
16:48engine has no UI. This is just a
16:50placeholder UI so you can see it. See it
16:53just logs things very bare bones. But
16:56then I select an equipment
16:58and you see this is for free because
17:01these arrows are already implemented in
17:03the engine. So I just say to the AI when
17:06I want to build an equipment
17:09choose the push direction and highlight
17:12it. So you push it like this. For
17:14example, I will build a shaft here. It
17:16costs incrementally more but I have $1
17:19trillion anyway. So what you do is for
17:21example you push another one. Okay. And
17:25then you push another one and it merges.
17:26It's a merge incremental
17:29right. But I had to deal with some
17:31limitations myself. So for example if I
17:35hope they are still here I will to show
17:37them to you external images. So you see
17:41I had a merge problem.
17:44Not that one. Shaft problem. So I just
17:47took these. But where is that? Oh, I I
17:50deleted them. So there was a graphics
17:52problem. Very basic. When I zoomed out,
17:56these lines, see these lines here would
17:59disappear, right? Because of antalyzing
18:02and such. Now Godo, Unity, Unreal, or
18:06pretty much any other engine you will
18:07deal with already has this. My engine
18:11Phaser did not. and my engine did not
18:15right. So I had to work with AI with
18:18cloud and it took about three or four
18:21days to handle it properly.
18:25I I had to work with AI to create for
18:27example atlases of sprites here because
18:31I handled that problem. These these
18:34lines would disappear. I wish the image
18:36was there to show. Wait wait wait a
18:38second. Maybe it lives in my screenshots
18:41folder. Let me try to find it.
18:46Ah, it's on the other computer. Not
18:47here. I work on the other computer.
18:48Anyway, no worries. But the thing is
18:52then I I solved these problems, these
18:55lines disappearing. But then I had
18:59complete fault lines like this because
19:02the map is being divided into chunks and
19:04rendered that way. So I had to solve
19:07that one. Unity has that built in.
19:10Unreal is not that good at 2D, but it
19:13has it built in. Godo definitely has it
19:16built in. And now my engine has that
19:19built in. But I did this because, for
19:22example, this game was not in this. This
19:25game was not an idea. This game was just
19:28because I didn't want to make something
19:30as complicated as this for a starting
19:33game in this one engine. Right? So I
19:36will build this one. But then I will
19:38have a hex engine. And actually when I
19:41talked to cloud as when creating this I
19:44said this will be this engine will be
19:46hex for now but in the future it will
19:50definitely have square grids as well
19:53rectangular grids. So plan for that.
19:57Does it make sense?
20:01So I don't know what it did inside. I
20:03admit because as I said I'm not a
20:05JavaScript coder. I don't know what's
20:07going on inside this grid. See this
20:10entire grid is the engine and this game
20:13is what you see on the screen. This is
20:15the game. Okay, it doesn't even have
20:18proper folder structure because I just
20:20let it do its thing. Maybe we should do
20:23this sometime. But this thing animation
20:26and I worked in modules animation module
20:29camera module clustering it. It detects
20:32clusters for biomes etc. for example and
20:35vis there is visibility some if
20:37something is
20:40further front it overlaps the it
20:46closes the thing on the back I'm a
20:50little bit sleepless and hungry I'm
20:52sorry about that but there was also one
20:56scene
20:58okay it public levels
21:01Round world 20. Let's try that one.
21:11But how do I round
21:1520? Let's try that. No, doesn't work.
21:19Hexagon pointy.
21:26No, you lied to me.
21:30scene. Demo demo scene. Ah, level. Okay,
21:34not scene. Sorry about that.
21:39But there was a tilted
21:47tilted
21:49demo level.
21:52So because this is top down but this is
21:55not right. You see it at an angle. So I
21:58said I told cloud to that this should be
22:02able to work at an angle and all the
22:05because it is 2D it cannot actually turn
22:08it. What it does is squish the tiles to
22:11make it look like it is turning right.
22:16So let's see where that one is.
22:20and hexagon tilted a copy and
22:24okay let's do what it says
22:29I think I deleted that out of it
22:32let's go into the level
22:35and it says add tilt
22:43so it's been a while so I forget about
22:46these things.
22:52No. Is it?
22:59Let's try. I mean, what can we lose? We
23:02are doing things together, right?
23:08No, didn't work.
23:12Oh, wait a second. Let's see. Where is
23:15the tilt value in this one?
23:20Ah, tilt is just written like that.
23:21Okay, that's why it didn't work
23:29like this. Let's see. I don't This is
23:33JSON. I'm not very proficient with JSON
23:35file format either because I never
23:37directly interface with it. But I can
23:40find my way around it. Tilt is equal to
23:43zero. And is it
23:450.5. Okay, let's try this again. Now
23:50see, so I can tilt it. But for example,
23:54right now I cannot rotate it any more
23:57than it is like this. But this is just
23:59things squished. So let's try that on
24:02the demo scene. For example, where is
24:04the demo scene?
24:09It's not in the levels than it is in
24:13source
24:15game demo scene. But this is
24:20yeah this is not the this is all this
24:22this is all the game so I cannot tilt it
24:24from here I guess. Anyway, but you get
24:27the point, right?
24:33We have spent the first half of our
24:36lecture on engines and how to build one.
24:41But I think it was fruitful
24:44and you can rest in bliss knowing that I
24:49will not ask anything from this in the
24:51exam. Probably not.
24:56because that come on. Yes, I will
24:59definitely stop using this.
25:06I mean, it's partially the software's
25:08fault and partially this thing is acting
25:10up. So, I hate switching from stylus to
25:14mouse, but for you guys,
25:20what was that for you? Nothing is
25:25It's a Fallout 4 reference and I cannot
25:27click on it now.
25:31Ah, come on.
25:38Okay, I'm right about to end this. Oh,
25:41you don't see my You don't see the
25:42struggle I'm in now. You will see.
25:46See? Do you see it?
25:53See,
25:55this is what I'm dealing with every day
25:57in my life. No, not actually. But why is
26:00it stuck?
26:05Okay, I'm officially done with this. I'm
26:08just going to do it with my mouse first.
26:11And I think it will stop acting up once
26:13I start type writing with it. No. Why
26:17does it still
26:19I think there's a button that has been
26:24okay guys three minute break until I fix
26:27this thing. Okay, let's take a five.
26:29Let's make it 5 minutes. It's 11. Let's
26:31meet 5 minutes later at 11.0 11:05 and
26:36this thing will be fixed. Okay, see you.
26:41Okay, welcome back everyone. Welcome
26:43back everyone. So
26:47before the break we discussed about what
26:49is an engine and how can we make one
26:52using an AI and then begs the question
26:55do we need to make one? Could I have
26:57made this thing that I showed you in
27:00Unity? I mean I could have no problems.
27:03Could I have used AI to make it in
27:05Unity? Yes, I could have. Same as in
27:08Godo. Why did I choose something like
27:10Phaser that I don't know how to write
27:13in? Right? Because for example for this
27:16specific thing,
27:19Unity is an editor first engine. There
27:22are many things that are happening in
27:24the editor and you have to see them. And
27:26as I said back in back in the day, my
27:29experience with Unity's MCP's AI with
27:31Unity wasn't very good. And also
27:34whenever I did something in Unity, I was
27:37going to have to go to Unity, press play
27:40button, check things, see if it works,
27:42blah blah blah. Whereas in here, I just
27:45interface with the AI and it
27:50it automatically puts the thing on the
27:52screen next to me when I type it. And I
27:55type it, I give it a command, it does
27:58its thing for three, four, five minutes,
28:00whatever, and I can do something else in
28:02the meantime. It's I'm delegating that
28:04work to the AI. And that is frankly the
28:07only way I can find time in my schedule
28:11to make a game right now. So
28:14Unity would require me to constantly
28:17step in adjust something in the UI in
28:20the inspector, adjust something in the
28:22settings blah blah. I didn't want to go
28:24through with that. So I wanted something
28:26headless. Headless means it works in
28:29code all by itself. Godo has such a
28:31mode. So we can try godo too. Anyway,
28:34now we are back to the topic at hand.
28:37So, let's go. Let's very carefully try
28:42the ah
28:45because this isn't in focus. It should
28:47be in focus. Yeah, here it is. A working
28:53stylus is a happy stylus. So, what did
28:55we talk about last week? We we talked
28:58about the steps of production. If you
29:00remember, we had the ideation stage.
29:07We had the pre-production,
29:17we have production,
29:23and we had post production. I don't have
29:27there's a USB problem here because my
29:30numpad micropad also flashes the lights
29:33which means there is a USB communication
29:36problem. Maybe the mainboard there are
29:38lots of USB devices plugged into this
29:40thing. So maybe I should make it powered
29:42postp production.
29:49Okay, so we already covered this one.
29:55So let's put double ticks on that. And
29:58now we want to cover this one.
30:03What is pre-production?
30:06What is the
30:10pre production stage? What do you think
30:13we do there?
30:23As much as I love the extra monitor, I
30:25think I'm going to go back to my old
30:28tablet.
30:29What do we do? What is your opinion?
30:32What do you think we do in
30:34pre-production stage?
30:36>> We start planning.
30:39>> Okay, good.
30:40>> My we start planning how the game is
30:43going to work. Uh the main game loop.
30:46>> Mhm. And
30:48>> how are we going to go about making the
30:51game? Like there's going to be different
30:54enemy type. So we make a default enemy
30:56type and change a little thing about it.
31:00>> Wonderful. Wonderful remark. You said we
31:02make a default enemy type. Where do we
31:05make it?
31:08>> Oh, that's also pre-production. I guess
31:10you choose which engine you are going to
31:12use or if you are going to use an engine
31:14and stuff. So because when you say we
31:17make a default enemy type it implies you
31:20code it in right but do you have your
31:23basic game mechanics laid down yet? The
31:27more important thing is have you decided
31:30what exactly your mechanics will be? Now
31:34there are different there is a
31:35difference between for example saying
31:38let's say
31:42and
31:44FPS game
31:48where
31:51we
31:54fight zombies
32:00and the name FP the genre FPS PS gives
32:03me a lot of information. I'm going to
32:05have guns. I'm going to shoot with them.
32:07Blah blah blah. Right? But these are the
32:11core mechanics. And now when I'm in
32:14pre-production, I want to decide maybe
32:16for example
32:20zombie
32:28limbs fly off.
32:35when you shoot them.
32:43Okay, there is definitely So, but
32:49this is an additional mechanic. It's a
32:51polish mechanic, but this is something
32:53you want to implement into the game
32:55because you want it to be appealing, for
32:57example, or maybe a more concrete
33:01mechanic
33:10you cannot get.
33:14Sorry, this is this is a total disaster.
33:21I will just type it.
33:26Huh?
33:30Wait a second. Why is my control key is
33:32pressed for some reason?
33:35Come on.
33:37Is it the other keyboard? There are two
33:39keyboards.
33:41No.
33:45Okay, this is coming to you live.
33:48And can I delete it? I cannot.
33:55What does this button do? It carries.
33:56What does this one do? It's scales. I
33:59like this software less and less. And I
34:02don't think it's the software's fault.
34:06Let's say the mechanic is
34:10you
34:13cannot
34:17heal.
34:21When you die,
34:24you restart.
34:29These may not be the things you have
34:30decided during the ideation stage. In
34:33the ideation stage, you wanted to make a
34:35cool FPS game where you shoot zombies.
34:37But now that you go into pre-production,
34:40you have to think of all of these things
34:42and you have to design for them, right?
34:46So in the pre-production unless we are
34:49doing something that requires it we are
34:52not touching the code we are not
34:53touching the engine we are planning okay
35:11we are not making the game we haven't
35:12even started making the game so what do
35:15we do
35:19I will start I will go in random order.
35:23So for example
35:25we will define our art style.
35:32Now you can say
35:35but we have we should have already
35:37decided that in the ideation stage
35:38because it's a very central part of it
35:40and you are correct but you may maybe
35:43you don't define
35:48But you refine the art style.
35:51You wanted handp painted graphics like
35:54Warcraft, but then you decide, oh, maybe
35:57we should go this specific route down
35:59it, etc.
36:07Okay,
36:09I will just remove this. I think this
36:11causes all my problems. game pad
36:13features and such. We will see. So, we
36:16will define or refine our art style,
36:19right? And
36:22we will decide.
36:27No, it wasn't a problem.
36:30It was just a symptom pipeline.
36:38Guys, can you give me a minute?
36:41I will do something. And I think I know
36:43what's going on.
37:29I'll be back. Let's see.
37:36No, it doesn't work at all. It's
37:38blended.
37:40No, it does.
37:46No, the same problems, the same
37:47everything. There's some interference in
37:49the USB ports.
37:51Maybe something is short circuiting. I
37:53don't know. I tried I switched the USB
37:55ports, but it didn't help.
37:58This is our problem today. So but we are
38:01planning our pipeline. For example, one
38:03very important thing that you should do
38:05first and foremost team formation.
38:10Who will be in which role?
38:25Right? Because you are six people and
38:28you should ple me this even if you say
38:33oh we are just four people everyone will
38:35do whatever they can you still should
38:39have your roles set up to a point
38:43because if there are three coders
38:47and each one is trying to pull the code
38:49base in a different direction you are
38:51going to tear that team apart pretty
38:53fast. You should know okay we are three
38:55coders but we should decide
38:59which who decides which code about what.
39:02Maybe one person will be specifically
39:04dealing with the combat code right the
39:07other will be dealing with the other
39:09part of the code and the third person
39:11will just do the pipeline tooling for
39:14the 3D artist. I don't know
39:18but that is what you're going to do and
39:21we're going to do something that again
39:23everybody despises documentation.
39:35Now
39:37this is important
39:40because I just told you documentation is
39:43the memory of the project.
39:55Imagine that you have a mechanic right
39:58in your mind. So this is the mechanic
40:02and you try started testing it inside
40:05the game and it didn't go very well. So
40:09you scrapped it. You said we won't be
40:11using this mechanic. So your game is now
40:13without that mechanic, right? And two
40:16months, three months pass, maybe the
40:19person who coded this or the designer
40:21who worked with this is gone. Maybe
40:22you're a large studio and you're a small
40:25studio and that guy who just worked on
40:28that left and a new guy came and just
40:31said, "Oh, why don't we introduce a
40:32mechanic like this?"
40:34Now, if you don't have any
40:35documentation, you might say, "Oh,
40:37that's a great idea." But if you do have
40:39the documentation, if you know this was
40:42scrapped and better, if you know why it
40:45was scrapped, you say, "All right, we
40:47tried this once. It didn't work, so
40:51let's not use it again." And please
40:54don't trust your memory saying, "Oh, we
40:58would remember." No, you wouldn't. Okay,
41:02this is especially valid about code. You
41:05do things a certain way, it doesn't
41:08work. You scrap it, you start over, you
41:11do something else. And then some other
41:13coder or some friend of yours comes in
41:16and suggests why don't you use this
41:18method and you don't trust me, you
41:21probably won't remember it. You tried it
41:24and it didn't work because it will be
41:26suggested to you in a different way.
41:29Right? So documentation is really
41:33important. It's the memory of the
41:36project
41:38and as always the memory is not a fixed
41:41thing. So the
41:46the main document we have in game
41:48development is called game design
41:55document.
42:00Now remember last week we talked about
42:02the high concept document, right? High
42:05concept document is the two or three
42:08page five minute read summary of the
42:11game. This is
42:14everything. So this is the main document
42:27about
42:29everything in the game.
42:44Let me take it down here.
42:51I'm very disappointed with this.
42:56Maybe it's the holding angle. Maybe it's
42:57because I'm not holding it upright. We
43:00will see. Let's try another thing. So,
43:02but the other important thing is guys,
43:05this is very important.
43:08It is a living document.
43:19which means
43:26constantly updated
43:32throughout
43:39the game's development.
43:42So you don't create this document and
43:44say this is the fixed document let's
43:46make according to this because you are
43:48going to change designs the certain
43:51designs you had worked perfectly on
43:54paper will not work at all right
44:07this is a living document ment.
44:13Why is it living? So it is not a fixed
44:16document that you design once, you write
44:19once and everything else is made
44:22according to it. Now
44:24you are going to put
44:27for example
44:29mechanics into this
44:34your level design philosophies. For
44:36example,
44:39level design rules. You will say the
44:42levels will have these rules at minimum
44:44and the level designer will look at
44:46this, right? And we'll design that. You
44:48will have your character design rules.
45:03All sorts of stuff. everything.
45:13So this can be a document that is
45:15hundreds of pages long, right? Okay. So
45:19but you're going to create start
45:21creating it here. But you will know that
45:24it will start changing. Okay. And for
45:28example
45:31you can start the most important part
45:34here is this is where you touch the
45:35engine for example prototyping.
45:45Now there are different types of
45:46prototyping. The cheapest one actually
45:51is one that you probably don't know
45:53about and you probably will never do
45:55even if I tell you about it. It is
45:57something called paper prototyping.
46:11And paper prototyping is exactly what it
46:14sounds like. You take a piece of paper.
46:17You take Lego pieces, whatever to
46:19represent your units. Okay? You take
46:22whatever is handy in the physical space
46:25around you. You use things like this,
46:27for example. You use
46:32things like this, right? With symbols on
46:35it, for example, as board pieces. And
46:37you just
46:40put things. Let me show how can I show
46:42it to you? Can I do this? I apparently
46:46can. Back to the future is here. Okay.
46:49So, you will just take these and you
46:51will say these are my tiles.
46:55Okay. And you will say
46:59now this unit looks here and this one
47:02will be here. So, if it's two steps, you
47:04can hit that. Let me see if it can hit
47:06that. All right. It could hit it. You
47:08can see my You can see here, right? It's
47:10a little bit too far like this.
47:13Let's make it No, I cannot make it any
47:16closer because
47:19you can understand what I'm doing here.
47:21I'm just taking things that I have
47:23around me or I can I make them with
47:25paper and pen and I say, "Oh, there is a
47:28barrier here now. So, this one cannot
47:30hit this one. So, it will try to hit
47:33this one. Can it hit?" But what? So, I'm
47:35testing out my mechanics on the board in
47:39front of me. Why do I do this and not go
47:42into Unity and start typing away?
47:44Because this is way cheaper not in
47:48materials. This is cheaper in time.
47:51Okay? When you do something like this,
47:55it will be multiple times cheaper to do
47:59because I just had these lying around
48:00from other things, right? But uh you can
48:04just cut pieces of paper, write letters
48:06on them, put them on and you can test if
48:09what you are trying to do has some logic
48:13in what you are trying to do. If it has
48:15some fun in it and I'm going to show you
48:18something. Let's see if we can get back
48:19here now.
48:22Okay, minimal damage. Wonderful. I'm
48:25going to show you something really
48:28important.
48:30Let's share [snorts] the screen.
48:37Let me just gather these Lego pieces
48:39back because they're getting in my way
48:42of my typing.
48:45Me put everything in place
48:48while you watch a wonderful sunset. So,
48:52I'm going to show you.
48:58It's a terrible video quality, but
49:04I think we will find that. So, this guy,
49:07I think you know him. He's kind of a
49:09famous level designer, right? And when
49:13he designed
49:16Metal Gear Solid one,
49:19he used this
49:21method. For example, he was trying to
49:24understand this is basically paper
49:25prototyping but with Legos and a webcam
49:29but he was trying to visualize how
49:33different viewpoints would look before
49:37even coding in something and
49:41in our defense today it is actually easy
49:43to do this in Unity etc. But I'm trying
49:46to put the point across. Back in his
49:49time, it was very difficult because you
49:51didn't have the engine, right? You
49:53should you had to write the engine to
49:56support all of these.
49:58[music]
50:00>> So this is the game view. This is a
50:03summary of what they are doing. And I
50:05said terrible quality, but this is top
50:08down view. And he shows with the camera
50:10what it's going to look like.
50:12And this is when you are investigating
50:14your surroundings. It switches the first
50:16person view and he shows with the camera
50:19what it's going to look like.
50:22[music]
50:33like this.
50:35[music]
50:39But you can see even if you have Unity,
50:44having a programmer move this camera
50:46around is much slower than just taking
50:49the webcam in your hand and poking
50:51around in a Lego set to see what
50:54different camera angles will show you,
50:56right? So this thing still has merit
50:59actually when you think about because
51:00see how he very smoothly moves for
51:03example.
51:07>> So he can actually even simulate camera
51:10movements with his webcam in Unity.
51:13Again it's going to take some coding and
51:15some moving things around and then
51:16pressing play to see all of this. But
51:18he's using a live webcam shoot.
51:20>> So you go up. So for I'll just shut it
51:23off. And then he says as you come up the
51:26camera will move here quickly. So he can
51:28do it right because you see in the game
51:31and after you come out it goes into
51:32another mode again. This is switching
51:37and
51:40while we're at it.
51:43So this is for app design but the same
51:46thing goes for app design as well. So
51:49you can just paper prototype things,
51:52right? And show how things are going to
51:54go. And I admit this is from a time when
51:57the AI was not here. So
52:04we can just say
52:07AI agent prototyping
52:16prototyping prototyper
52:24we can say this right. So but we also
52:29have our traditional
52:31uh digital prototype.
52:34So what we do
52:46and the common of the feature of this
52:49one and now of this one as well. These
52:52would probably be disposable.
53:00What I mean is you will write quick, you
53:03will write dirty, you will cut corners,
53:06take shortcuts, and this prototype will
53:10probably not work in production because
53:13it will be too complicated to put
53:16something on top of this. You just want
53:18a simple WD move set and the basic
53:21camera movements. So you will move the C
53:23character around and see if what you are
53:25doing, put the basic mechanics in place
53:28and try to see what you are doing is
53:29fun. And if it is fun, you then sit down
53:33with all your game design constraints
53:36and you try to figure out how you will
53:38actually do it. As I say, AI has sped
53:43this up immensely.
53:46This is where you would see the true
53:48value of working with AI. Okay? Because
53:52you can prototype.
53:54For example, you are going to work in
53:56Godo or Unity in 3D, but you can just
53:59use an AI agent to prototype it in
54:02JavaScript in a 2D mode from top down
54:06with just a few prompts and a few hours
54:08of working on it, right? Whereas if you
54:12want to do it in Unity as a prototype,
54:13maybe you will spend two or three days
54:15to put everything in a very crude way
54:18together and see if it works. But you
54:20can just say build me a prototype of
54:23this and I have these ideas blah blah
54:26blah and it can create a structure for
54:28you. You can even do that structurally.
54:30It can create you a good coherent
54:32structure and build it and you play and
54:35you see if it is fun or not and then you
54:37say okay this is quite fun. So let's
54:40start building it for real. Does it make
54:43sense?
54:47This step is really important. Thus, and
54:51I can tell you
54:55pretty much all the games I made in my
54:58career were prototypes. Almost all of
55:02them because I worked in the mobile
55:04industry. And in mobile, you do this.
55:07you make the prototype, you do you put a
55:10you don't work on it too much. So
55:14uh when I was in come on
55:26when I started working on mobile
55:28industry the thing we called hyper
55:30casual was made in about three
55:34weeks or let's say two and a half weeks.
55:41So developing one prototype took about
55:43yeah it took about three weeks but no
55:46that was my first one because I was a
55:49noob back then. So it took about two and
55:52a half weeksish
55:56right when I left it the mobile mobile
56:00not exactly when I left hyper casual
56:02let's say not mobile completely it took
56:05about 4 days
56:09and this is a span of what 3 years four
56:13years
56:15so things and there was no AI back then
56:18no chat GPT wasn't even a thing. Okay,
56:22maybe Google even didn't publish the
56:24paper. No, Google did publish the paper,
56:26but it wasn't it was being developed in
56:28secret. So, but in three or three to
56:31four years, prototyping went from in two
56:35and a half weeks to four days. This is
56:38an almost a game that almost looks like
56:42the complete one in order to test it on
56:45people, right? So,
56:48and I made I think around a hundred of
56:51these
56:53and
56:5595 96% of them did not ship. One of the
56:58games I worked on was shipped but it
57:00wasn't hyper casual. That was later on.
57:02I think none of the hyper casuals I made
57:04shipped because in Turkey back in those
57:07days maybe they were releasing thousands
57:10of hyper casuals
57:13per month. Overall all the studios in
57:15Turkey, Roli alone used to work with
57:19around 90 game studios in Turkey and in
57:26Balkans I guess. So they had over 90
57:30game studios working with them and each
57:33of them was pumping around 10 apparently
57:36or there are others publishers as well
57:38but at least 1,000 games were hitting
57:40the market every month being tested
57:44every month. Okay. So only one or two of
57:47these are hits.
57:50So it's a
57:530.1% or like this chance to have a hit.
57:58So I never had one.
58:01But this is hyper casual. There's also
58:04hybrid casual, there's strategy, there's
58:05blood, there's premium, all that things.
58:07But that's not necessary. What we are,
58:09what I'm trying to tell you is you make
58:12a prototype
58:14to see what you are building is actually
58:17fun, if it actually works. Because if it
58:20not, it's not you are still in the early
58:22stages. You can cut it. Go back to
58:25ideiation and design something else. All
58:27right?
58:30Is this
58:33okay? Did you understand what I'm trying
58:35to do here? You do your prototype in the
58:37pre-production stage. And generally,
58:42let's put a yellow arrow, blue arrow
58:44here.
58:53Generally
58:55the prototype
59:07has to be
59:11completely rebuilt.
59:24for the game
59:31because the prototype is just to see if
59:33it will work. You just hack and cut
59:36corners and tie everything with duct
59:38tape. Okay, tie them together. But when
59:42the game gets larger, it generally does
59:45not hold. Well, now what people do
59:47actually
59:49is they is they build the prototype.
59:54This is the prototype, right? And they
59:57tell themselves, okay, when we decide
1:00:00that we going to build this, we will
1:00:02scrap this and do this proper. But then
1:00:05since all those systems seem to be
1:00:07working, they just start building around
1:00:09it anyway.
1:00:16like this. Okay, it's not a Microsoft
1:00:19logo. So in the end, what happens is and
1:00:24if you also don't have proper
1:00:26documentation, this is an abomination.
1:00:29Your game is an abomination that only
1:00:32you can understand.
1:00:35It will be a nightmare for any other
1:00:37coder you will bring in. And even if you
1:00:41even you two months from now will not be
1:00:44able to understand this if you just
1:00:46leave it and come back to it. You will
1:00:48curse at you in the past for doing it
1:00:52like this. But it is the easy path
1:00:55because you already have something
1:00:56working. So why not just keep building
1:00:59on it? Keep it in mind. You can do it in
1:01:02a structured way from the beginning of
1:01:05course but you have to keep these things
1:01:07in mind. All right. So, and the other
1:01:10thing that we have to do the last thing
1:01:15did I write pipeline planning here.
1:01:19Okay, let's just scroll with the mouse.
1:01:22No.
1:01:25Yes, we did the pipeline planning.
1:01:28But we have to do our for example
1:01:30technical tests.
1:01:40Because we have something in mind. Is it
1:01:42working? Can we do this?
1:02:04This is just an example.
1:02:06Sometimes you're going to test for a
1:02:08visual style. Can I achieve that visual
1:02:10style? Actually, is it possible? I have
1:02:12something in mind. I want things to look
1:02:14like a one log painting. Well, because
1:02:18the answer is
1:02:21can we do that in gaming? The answer is
1:02:25yes.
1:02:27Because you are just printing pixels on
1:02:29the screen. We are not doing anything
1:02:30else. So if you print if you find a way
1:02:33to print the correct pixels on the
1:02:36correct position on the screen, you can
1:02:38do anything that is defined by 2D
1:02:40surfaces. The point is can you do that?
1:02:45Right? The question is not can it be
1:02:47done because the answer is probably yes
1:02:51but the the true question is can we do
1:02:54that mechanic in this engine with
1:03:00now it comes with
1:03:05the money
1:03:09time
1:03:12or talent that's not like talent it's It
1:03:16dehumanizes people a little bit. People
1:03:20that we have
1:03:26now this is probably the most important.
1:03:30I wrote technical tests but I will erase
1:03:34this because there's a much more
1:03:36important word for this because it's not
1:03:39about technical at all. This is
1:03:41something behemoth called and tell me if
1:03:45you have heard it before
1:03:51Scope
1:03:53because maybe you do want to make an
1:03:56openw world
1:03:58sandbox survival crafting game with deck
1:04:01building mechanics and rogike
1:04:04progression. Can you do all that with
1:04:07what you have? You're going to you are
1:04:09going to look at your money me okay it's
1:04:11just me and my computer I don't need the
1:04:13money fine you don't have a money
1:04:15constraint wonderful
1:04:18and it's only me all right do you or my
1:04:22friend who does 3D okay do do you two
1:04:26have the capability then to do this
1:04:29and most importantly do you plan to
1:04:32release this anytime soon because if you
1:04:34give yourself 40 years to develop a game
1:04:37you can pretty much do anything right?
1:04:40So this is
1:04:44define scope
1:04:48is the most important thing you will do
1:04:52in this stage. You will say okay this
1:04:54will be an action RPG or this will be an
1:04:57zombie FPS so we probably don't need
1:05:01crafting in it.
1:05:03I mean can do we want crafting? Yes, we
1:05:06actually do. But we want to make this
1:05:09game playable in one year, for example.
1:05:11And if we put all these systems in, the
1:05:14crafting will be quite a lot of work. So
1:05:17you can say, all right, I'd say what I
1:05:19have my core features,
1:05:24right? I definitely want in the game.
1:05:27And then there are nice toes.
1:05:38And then there will be some
1:05:41that will be that will look very good on
1:05:44paper
1:05:47but that will be bloat. So you don't
1:05:50actually want them right. So you have to
1:05:53define your scope carefully.
1:05:58And the let me tell you if I can zoom
1:06:03out of this thing.
1:06:08So everything I said here see look this
1:06:12is this is the maximum zoom. All right.
1:06:15So you see production pre-production
1:06:17here we talked a lot about about a lot
1:06:19of things. Two things two two most three
1:06:23everything is important but the most
1:06:26important thing here is the scope.
1:06:29You have to be aware what you want to
1:06:33have in your game. Okay. What you want
1:06:37but cannot do.
1:06:40That's reasonable. what you want but
1:06:44wouldn't be fun in this game or would
1:06:47ruin the experience. Right? You have to
1:06:51find these things and you have to look
1:06:53at them. The way to
1:06:56refine your scope
1:06:59is to prototype.
1:07:02You refine by prototyping and testing
1:07:05things.
1:07:07And you write things in your game design
1:07:10document.
1:07:12Now you will think okay at this point I
1:07:15am making the game actually right
1:07:17because I I'm defining my scope I'm
1:07:20prototyping I'm writing a game design
1:07:22document but
1:07:25imagine that you are going to make a
1:07:28game that will take about three years to
1:07:30complete that's your vision right
1:07:32because we are four people we are also
1:07:34working we are not doing this full-time
1:07:36or even if we are doing this full-time
1:07:37making games is a long making a quality
1:07:40game can be a long time. So
1:07:44spending two months
1:07:47on proper prototyping is not actually
1:07:50wasted time because if the game doesn't
1:07:53work properly, it it was wonderful in
1:07:56your head. The ideation stage went
1:07:58perfectly. You said, "Okay, we're going
1:08:00to do this. The character will fly, will
1:08:01punch monsters, there will be zombies,
1:08:03there can even be dinosaurs, aliens. Oh,
1:08:05what? Wait, let's put crafting in it."
1:08:06You are ideating. You are throwing
1:08:08everything around, right?
1:08:10and you have the perfect thing in your
1:08:13head and then you sit down to actually
1:08:16making it. Sometimes, most of the times
1:08:19you will realize, oh, it's not as fun as
1:08:22we thought it would be.
1:08:25Then you don't lose one year, two years
1:08:28to realize that just by going into
1:08:31production. You realize that in two
1:08:33months. Yes, you lost two months, but at
1:08:36least you didn't lose a year or two.
1:08:39Okay, does it make sense?
1:08:43Any questions?
1:08:48Then I will do a happy dance because you
1:08:51all got this perfectly and there are no
1:08:54questions at all. So we couldn't go into
1:08:57production and post-production. We are a
1:08:58little bit behind. But hey, we talked
1:09:01about the engines and I think how to use
1:09:03AI with the engines and I think that is
1:09:06important because
1:09:09you will hopefully go into game design
1:09:11and development and you will want to use
1:09:13AI with it naturally so and you should
1:09:15it's a great tool right now to be doing
1:09:17things and
1:09:20it is important that you know how to
1:09:23properly use it to get work done. Now
1:09:28there is an important distinction here
1:09:30please. I want to make my camera I mean
1:09:33ever since I showed you that my camera
1:09:35is a little bit off. I think this should
1:09:37be the correct one. Okay. So
1:09:40here is the important thing. Now
1:09:45this semester I will keep talking a lot
1:09:47about AI because it is here. It is here
1:09:50even more than it was last year. It is
1:09:53here even more than it was last
1:09:55semester. All right. It is more here now
1:09:58than it was last week. The thing is
1:10:01improving people. Have you seen that
1:10:03people are putting Mario in Elden Ring,
1:10:05right? Because now the latest models
1:10:08Astra and Fable can reverse engineer
1:10:12game code and inject another game's code
1:10:15or rewrite another game's code in that
1:10:18game's code.
1:10:20So this is something that has that could
1:10:22never be actually there's something I
1:10:24will show you a website. Was it? Let me
1:10:28just just a second.
1:10:31I have to find it first.
1:10:34Some company reverse engineered
1:10:36Photoshop
1:10:41and
1:10:43wrote it again. Now, this is not
1:10:46illegal.
1:10:48They basically wrote another copy of
1:10:51Photoshop, right? Let me share my screen
1:10:53with you.
1:10:59This is let me
1:11:02show put the link in the chat as well
1:11:05for those of you who want to explore
1:11:08that. So they made Photoshop they made
1:11:12what was that in design? No, what was
1:11:15Photoshop? What was Adobe's vector
1:11:17program? Illustrator.
1:11:20So they made
1:11:23Premiere Pro, Lightroom,
1:11:28Acrobat. I will use this a lot.
1:11:32F what was Microsoft Adobe's? Do you
1:11:36know what the After Effects
1:11:38and I don't publisher I think I think
1:11:41but publisher was decry whatever anyway.
1:11:44So, but these guys reverse engineered
1:11:46these and this is actually the same
1:11:51interface as Photoshop. The same
1:11:54shortcuts, the same workflows, the same
1:11:56everything. Because up until now, the
1:12:00real contender was
1:12:04And the way Linux people do it is
1:12:07generally completely redesign everything
1:12:11in their own mind. or we had photo
1:12:15which is quite great by the way I still
1:12:17use it daily. This is a very good
1:12:19Photoshop clone but it lives on the web.
1:12:22So what these guys did is they reverse
1:12:25engineered it and I I suspect they
1:12:27heavily used AI for that right and they
1:12:31made a completely free Photoshop version
1:12:34and I will be using it quite a lot. So
1:12:39the use of AI is not even a discussion
1:12:42anymore. Well, if you talk about art,
1:12:45you can say that it is JI is a some is a
1:12:48topic different entirely. All right?
1:12:51Because that is about
1:12:54well copyrights and artist rights and
1:12:58everything. But in coding, in software
1:13:01development, the AI is here. Now you
1:13:06will ask me
1:13:08do we even need to learn coding anymore?
1:13:12That's and that would be a valid
1:13:14question because you have coding classes
1:13:15right and I say yes right now you do
1:13:20because
1:13:21the AI is not another way to code. You
1:13:27have to understand this. The AI is
1:13:30another agent another being not human
1:13:33but machine. It would be the same if you
1:13:36gave it to a human that does some coding
1:13:38for you, does the coding for you. So you
1:13:41have to understand the systems. You have
1:13:43to understand how the computer works,
1:13:46how code works to be able to direct AI
1:13:50properly. Otherwise, it will just go
1:13:53ahead and do what it thinks you want.
1:13:57Right?
1:13:58So even if it does it flawlessly, it may
1:14:02not be able to do what you want it to
1:14:04do. It might be doing something else.
1:14:06And you say, "Okay, then I will just
1:14:08prompt again." Well, it all costs money
1:14:10and tokens. You see, you go out to
1:14:13Twitter, formerly known as Twitter X,
1:14:16and there are AI bros there. Bro, I made
1:14:19Elen Ring in just three lines of prompt,
1:14:21blah blah blah. They just talk a lot,
1:14:24and you see them, and you say, "Oh, I
1:14:26must be missing something. These guys
1:14:27are doing so much with AI. Well, they
1:14:31look like they are doing so much with
1:14:32AI. What most of what they do is
1:14:34useless. They are playing around with it
1:14:36because they create content with it,
1:14:39right? And they don't tell you how much
1:14:41they spend in the background. They say,
1:14:44"Ah, Astra did this with just $60 of
1:14:46prompting." Okay, it is $60 for that,
1:14:49but what else did you do to acquire that
1:14:52prompt knowledge to tell it what to do?
1:14:55Have you spent $3,000 last week? You are
1:14:58not telling us that, right? So,
1:15:02long story short,
1:15:04coding is still relevant.
1:15:08Not about writing the actual code, but
1:15:11about knowing how code works. Okay.
1:15:16So, any questions now?
1:15:24Then I will leave you
1:15:27with a goodbye or with a Yeah. So I will
1:15:32first stop my this recording, then I
1:15:35will stop the Zoom recording. But at any
1:15:37rate, goodbye. Take care and go play
1:15:40some games.

About this transcript

This page contains the full transcript of ISU Fall 2026 - DGD205 Week 3 - 08.10.2026 by Onur EREREN, generated from the public captions YouTube serves with the video. The transcript has 9,193 words across 1,395 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 choose a caption track available on the video.

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, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.