Construct 3 in the AI agent era — what actually makes sense?

Not favoritedFavorited Favorited 2 favourites
From the Asset Store
Professional AI chatbot for Construct 3 with OpenAI API support, multiple models, conversation management, and full resp
  • Following Ashley reply, thats my exactly experience with AI and I think the only useful tool is the autocomplete and even so the impact is only 5% to little to be meaningful

  • but also i just want that for exemple i say to the ia to make a score system and the ia make a score system to the game. but i dont want any ia generated content.

    But that's the thing, that score system is still ai generated content, and steam and itch.io are requiring you to say if your game is ai generated or not- whether or not you used ai for the full game doesn't really matter.

    I honestly think it's just better to put in the effort of making it yourself, than risk backlash, or people not wanting to play your game, because of ai- not to mention the possibility of ai being copyright infringement in some way given how it's trained, which could also affect games and other media made using ai.

    just my opinion.

  • I think I exaggerated in my first post here about the importance of implementing AI in Construct 3. I was afraid that the technology would be ignored because of the hysteria of some people.

    I think Ashley's “wait and see” approach is ideal, and I'm glad that he is at least open-minded about the subject.

  • I'm making a fairly large game by engine standards. It's an isometric action/adventure game, with ~7,000 events at the moment.

    My second developer is also an artist, so I'm stuck working on my code alone. :)

    To avoid turning development into a never-ending process, I use everything I can, and as soon as AI Studio came out, I created a bunch of different assistants to understand how it works and what real help it can provide.

    Ultimately, I settled on an app that fetches a C3 project from a GitHub repository (it updates instantly, as it ignores everything except ES and scripts). I select the sheets to work on in a given session, and the model can also open related sheets for analysis. The AI ​​sees the sheet's JSON, and I'm the C3 editor. Save, push, and the assistant sees the changes in the project.

    The app lets you select a model, adjust the creativity temperature, and the context window bar displays the number of tokens used.

    Having ingested the project sheets, the model created a prompt for itself that allows its code to be pasted into the C3 editor, with varying success. This works well for modification, but less so for creating code from scratch. Yes, large models are proficient in languages, but valid C3 clipboard syntax was clearly not a priority in their training.

    On the one hand, it's annoying to wait for an assistant to correct its syntax, but on the other, it's a barrier to errors.

    So, in order of usefulness:

    1. Information. AI can do this instantly — for example, find Ashley's non-obvious forum post in which he admits that the animation trigger is delayed by one tick, and he himself doesn't remember why. Or the AI ​​will produce questionable information, which you criticize, thereby creating a logical conflict. Neural networks work better in conflict. So, if you're up for the investigation, you'll get a wealth of information for creating and releasing the game, step by step.

    2. Mathematics. The AI ​​rewrote all isometric distance calculations for detection and movement, with forward and reverse transformations. It wrote JS scripts that converted angles to radians, linked them to callable functions, and suggested using an angle matrix for massive trigonometric calculations—and even seemed to enjoy this part of the work.

    3. Debugging. As of late February 2026, two top-end models (Opus 4.6 and Gemini 3.1 Pro) were able to understand my project's logic and find errors in linked sheets of 500-1000 events 8 times out of 10. But they're very expensive.

    And there's the free Gemini 3 Flash, which can make guesses, 4-6 guesses per message. And if it doesn't guess, it will give you a sense of collaborative work, and you'll find the error on your own, but with less effort.

    4. Design. A special case of the first point. In fact, sometimes it's actually very useful. One good conversation can save you from a mistake. Almost all AI assistants are good at detailing and planning, which gives a more unbiased view of your project and its architecture.

    5. Generating code for game mechanics in the form of classic C3 editor events. This is the weakest point of almost any AI. It's not even a matter of syntax, but rather that the neural network always simplifies the code within the specifications, weakly analyzing its compatibility with the project.

    Judging by current progress, accessible models capable of responsible analysis should appear this year. This means that all programmers will use AI in one way or another.

    What can Scirra do? I honestly have no idea.

    Today I saw a video on Twitter of a large 3D game made based on a prompt, and a program that controls agents working on the project. Too much money has been invested in the tsunami heading our way.

  • envoys Wow, this is really cool! How do you set up a system like that? Do you have a link to a manual or a tutorial?

    Debugging. As of late February 2026, two top-end models (Opus 4.6 and Gemini 3.1 Pro) were able to understand my project's logic and find errors in linked sheets of 500-1000 events 8 times out of 10.

    Amazing! Do you need to explain every specific bug, or does it analyze the code on its own and look for potential issues?

    Today I saw a video on Twitter of a large 3D game made based on a prompt, and a program that controls agents working on the project. Too much money has been invested in the tsunami heading our way.

    Yep.. In a year or two we will be flooded with thousands of AI-generated games. Just like the internet is already flooded with AI-generated music and other content. I don’t understand how people don’t see this coming.

  • dop2000

    Yep.. In a year or two we will be flooded with thousands of AI-generated games. Just like the internet is already flooded with AI-generated music and other content. I don’t understand how people don’t see this coming.

    It's not coming it's already here in form of overwhelming amount of pure garbage, that's why people are against it.

    I honestly don't understand how you can look at this slop wave and say to yourself "yes, this is a good thing. More of that please".

    We all see the same thing, we all know what's coming. The difference between our opinions is not in the misunderstanding of the circumstances, but rather in the answer we give.

    My answer to this - games are not only products, they are also art and your code is an extension of the game's rules. An artist can't code review properly and I say it as an artist myself. To code review properly I have to go through the junior to mid levels of programming, which I won't be able to do if ai is doing junior to mid tasks! I wouldn't grow as a programmer I would grow as a vibe coder, who is not a good code reviewer.

    So how do I even reach the level of a senior+ if everyone keep telling me "just let the code monkeys do the job."? I'm in a position that I either accept the new rules and just disappear as human, or grind my coding skill so I can at least review. But no one is going to ever tell a junior "hold on on using ai. Develop your skills before you can review ai's code." No. Everyone and their cat is trying to sell ai to both juniors and seniors alike, while in reality it seems like it's beneficial only to a small minority of people at high positions and only in small runs.

    You said yourself that you need to be a senior+ dev to properly utilize agents(who in this context are essentially just digital vibe coders), but now it seems like you are fine with vibe coding the entire game.

    Where is the line between bad "vibe coding" an the good "proper use of ai"?

    I will add that the idea that you will be able to create a large scale project alone with vibe coding as your primary tool is just as unrealistic as before. The answer to that problem was always "scale down", because someone will be accountable for the result, ai generated or not, and that person has to be real and skilled.

  • but now it seems like you are fine with vibe coding the entire game.

    Where did you even get that from? I never said that and of course I'm against it. I said I need an AI assistant — something that helps me work faster, write cleaner code, catch more bugs.

    Just because some people use AI to generate low-effort slop doesn't mean we should avoid using it responsibly to make better games.

  • Try Construct 3

    Develop games in your browser. Powerful, performant & highly capable.

    Try Now Construct 3 users don't see these ads
  • If you choose to ignore AI, you’ll miss out on its benefits and see your work purely as art. Unfortunately, sooner or later, AI will inevitably replace much of this effort.

    We either adapt, or we choose to avoid it and let other game-making platforms take the lead. Construct 3 will be left behind if we cling to doing everything the traditional way.

  • > but now it seems like you are fine with vibe coding the entire game.

    Where did you even get that from? I never said that and of course I'm against it. I said I need an AI assistant — something that helps me work faster, write cleaner code, catch more bugs.

    Just because some people use AI to generate low-effort slop doesn't mean we should avoid using it responsibly to make better games.

    Sorry if I misunderstood your position, but you said: "In a year or two we will be flooded with thousands of AI-generated games. Just like the internet is already flooded with AI-generated music and other content." and it reads like you are fine with it and actively push for exactly this thing to happen. Ai music and other content is entirely ai generated and you are putting games in the same category, so that's what I assumed.

  • and it reads like you are fine with it and actively push for exactly this thing to happen

    I'm not fine with it, but there's nothing we can do to stop it. Accepting AI and using it in our work is not "giving in" - it's actually a way to stay competitive for a bit longer, before most of our jobs inevitably become redundant.

  • > and it reads like you are fine with it and actively push for exactly this thing to happen

    I'm not fine with it, but there's nothing we can do to stop it. Accepting AI and using it in our work is not "giving in" - it's actually a way to stay competitive for a bit longer, before most of our jobs inevitably become redundant.

    I see your point, but so far ai has been as much of a problem for everyone as it is a tool. I myself tried using it in my job - dropped it as soon as as I understood where it's going. My coworkers did, the result was not a success: company closed due to unrealistic timelines ai promised, people still overworked and underqualified.

    I understand that this is not a result of ai itself, but rather a bad implementation of ai, but that implementation was based on the same logic - "we'll be missing out if we don't jump on that train." So we jumped, and it sucked and we all felt no authorship or ownership and it solved zero issues. Essentially it just replaced all the pipelines with a mix of gacha and occasional manual fixes and everyone hated it, but everyone kept doing it, because reasons.

    It's not a guaranteed success, it's a bet.

    I think we will all be needed as long as ai keeps failing and oh boy it does. I also think it's important to understand that we're not really in a race. For end user, the amount of time spent on the code or art or music is irrelevant. For that small fracture of time the price is too high. All that somehow makes me more optimistic about ai. I think a lot of bad practices will be filtered in time. Maybe with some laws.

  • I hope most of you try vibe coding some ai slop. It really started clicking in Dec of 2025 and with the new models this month it is greatly improved. If you want to write games without coding it fits the bill and it is just so much fun.

    yours

    winkr7

  • envoys Wow, this is really cool! How do you set up a system like that? Do you have a link to a manual or a tutorial?

    > Debugging. As of late February 2026, two top-end models (Opus 4.6 and Gemini 3.1 Pro) were able to understand my project's logic and find errors in linked sheets of 500-1000 events 8 times out of 10.

    Amazing! Do you need to explain every specific bug, or does it analyze the code on its own and look for potential issues?

    > Today I saw a video on Twitter of a large 3D game made based on a prompt, and a program that controls agents working on the project. Too much money has been invested in the tsunami heading our way.

    Yep.. In a year or two we will be flooded with thousands of AI-generated games. Just like the internet is already flooded with AI-generated music and other content. I don’t understand how people don’t see this coming.

    Heh... that'll be a lot of text :)

    Just type in AI Studio: create a GitHub agent to work on the Construct 3 project.

    Currently, the top model Gemini 3.1Pro is building applications there, so feel free to communicate and describe your needs.

    So: on the left is a column with a public GitHub repository (open sheets with a click), in the middle is a chat with the model, and on the right is a window for the code that will generate the model, with a copy button, as well as tabs for open sheets from the repository. Plus: model selection from a drop-down list, with temperature selection, a "context window" scale for the model up to a million tokens, and a reset button. Select repository sheets before starting a session, attach images and files in the text input window, and get screenshots from the clipboard. The first message should start the session. The model should have access to the repository, open and view any sheets, just like an agent. Have an internal code validator for the C3 clipboard. The model should read the project's knowledge base and report this at the beginning of the session along with the greeting.

    (If desired, you can turn the model into a full-fledged agent and send its edits to GitHub, but that would be crazy now that the AI ​​understands C3 syntax so poorly.)

    When uploading the application, you need two text boxes: one for entering the repository address, and the other for the classic access token, which you can get in the GitHub settings. You can set this to be used by default if this project is your permanent place of work.

    You'll need to install GitHub Desktop to quickly push the project to the repository.

    After generating the application, you'll need to tell the coder what improvements need to be made and any bugs fixed. Be prepared to spend some time on all of this. My application saves the entire session in LS, so I can return to it even after closing the browser.

    Currently, the best time for free accounts in Europe is from morning until lunchtime, while America sleeps. The quota resets between 8 and 11 AM, depending on the time zone. Responses may degrade in the evening due to workload.

    First, upload as many system sheets as possible and request an assessment of the entire project. For work, select a specific problem and describe it; this works best.

    Create a knowledge base file for the AI ​​in the project and attach it to the application tabs so it loads automatically in each session. Describe your project and navigation principles (preferably through comments), and then add a prompt for code validation.

  • When the market will be flooded with AI generated products, people will be fired and the same people will not have a job, who will have money to buy your "cool game", especially since everyone will make their own once AI will advance more. So using AI is like committing suicide and bring down with you as many people as possible.

    Still many miss the ethical point, not the feasibility of using or not AI to make games or whatever. And I can still see that not even one has ever replied to my points, because AI users know that they can't provide a ethical answer to it.

    The fact that everyone will use AI then if you don't will be behind is the same like saying everyone is stealing/killing/committing violence, if you are not doing it you will be behind. Who cares about ethics? Just think about money, money that as said above there will not be there in the future. It will be like to try to sell air to breath in a clean world. Everyone can breath, why should they pay you money? And again, who pays the bills? Still no one answers to all those questions.

    Guys, this an ethical issue, not a technical issue. It's like the difference between developers that wants to make gambling games because who cares about other people's life, I need money and the ones that it's ok to make little money as long as I feel my game is providing happiness to people, not problems.

  • Still many miss the ethical point

    I'm not a fricking philosopher I'm here to make web games and farm moola from kids playing them

    Vibecoding go brrrrr and u go back to your english major classes or something

Jump to:
Active Users
There are 0 visitors browsing this topic (0 users and 0 guests)