Sangwright - browser-based vampire text game, looking for feedback

Playing Project Announcements
, , , ,
1

Howdy all! I’ve been reading through the forums and wanted to share something I’ve been building:

Sangwright is a choice-based text game set in a fictional city called Ashenmere. You play a new vampire across six nights, picking a district and an approach each night, with consequences that carry forward. Three districts, seven paths through the story, and a stat system (Blood, Shroud, Waning) that tracks what you spend and what you lose.

A few things upfront since I know the community asks:

What it is: Browser-based, mobile-friendly, no account needed, nothing to install. Each night takes about 3-5 minutes. It’s closer to a serialized choice-based narrative than a parser game; think Fallen London’s text-forward structure more than Zork.

How it’s built: I use Claude (AI) for the technical development: the React app, the backend, deployment. The prose is AI-assisted: I develop the world, the characters, the story, and the emotional beats, and collaborate with AI to write it, then edit by hand. Reference art is AI-generated. I’m tagging this ai because I’d rather be transparent from the start.

Where it came from: I played Vampire Wars (a GodWars MUD) too much in the late 90s. I’m a sales engineer, not a game developer - this is my first text game and I’m learning as I go. (thanks in advance for the grace!)

It’s free, no signup required: Sangwright.com

If you try it, I’d love to hear:

  • Does the writing pull you through, or does it stall somewhere?

  • Do choices feel consequential, or cosmetic?

  • Where did you stop, and why?

Happy to answer questions about how it’s built, the AI workflow, or anything else!

2

I’ve only had a chance to spend a few minutes with it but: it’s very pretty!

Is this a custom approach, or are you using any library/framework for the IF-y choice system?

1 Like
3

I gave this a try. I played through it twice, making different choices each time.

To answer your questions:

The writing is decent enough. I wouldn’t say it “stalled” as such, but the use of AI assistance is definitely evident to me. I’m not categorically against AI-generated prose or anything, but if I’m going to read something written by an AI, it’s usually going to be something that I generate myself to get the exact story that I want to read. So it’s not like it really grabbed me either.

Right now, the choices feel cosmetic. There are three stats that go up or down (or, in the case of waning, only up, I guess) but I don’t know what those numbers really change, if anything. Maybe this will matter more in the full game but right now I’m not sure what effect my choices even have on the story going forward.

I stopped after two playthroughs because I figured I’d seen enough for now, though I’d probably come back to it if you added more content.

I’ll also add that I like the idea of gradually and naturally introducing the lore of the world to the player rather than just info dumping it, but at least to me it was a little too casual and it felt a bit like the time I tried to play Vampire: the Masquerade without reading any of the books. There were entities or factions with various names (e.g., Kron, Skein, Coil) that apparently did different things and I would’ve liked a little more of an introduction as to what they were all about.

1 Like
4

Hello and welcome!

The following is a text dump that isn’t very organized, but is ordered chronologically:

DURING:
It definitely reads like obviously AI-generated prose to me. If someone were to claim it as completely human-generated, I would be skeptical enough to question them. The tropes are very cliché. Some of the animations felt unnecessarily slow, particularly the time transitions. Many details are told rather than shown—I hear about a woman I’ve “seen twice now” but I…don’t recall ever seeing her? “Your choices shape these numbers” seems rather obvious. Why is the climax called “the climax”? It’s strange to label your story beats.

AFTER:
Maybe it was my ending, but I didn’t feel like anything happened. I tried to play the sneaky spooky assassin vampire, since that’s how I like my combat, but that just ended up being anticlimactic. So for the third run, I played full Kron-aligned, but even that was quite predictable. I also didn’t know much about what was going on, like the different alliances and what they valued, what they believed, their histories, etc. Showing the location of each night seems unnecessary since I don’t have any control over where I go.

QUESTIONS:

  • I read quickly, but I found myself skimming through most of my runs. The fact that it’s big and italicized didn’t help. It’s body text, not a pullout quote.
  • Choices feel consequential but I don’t feel like I have much control over what’s going on. There’s a very clear “here are the results of your actions. Here is the buildup to the next choice. Here are your two options” setup that isn’t very compelling, especially when the choices are clearly directed towards an archetype.
  • I would’ve stopped after the first run (maybe even the first night—even if I didn’t know it was AI-generated, the writing is…not my style) but I kept going for the sake of useful feedback.
Final results

1
11092×1398 121 KB

2
21104×1326 118 KB

3
31080×1360 123 KB

4
41132×1412 125 KB

3 Likes
5

Thanks for taking the time to play through!

To answer your question: It’s a custom build, using React/Vite/Tailwind, no IF framework underneath. The choice system is just tracks which path you’re on and what you picked, then serves the next block of prose and stats accordingly. Nothing fancy under the hood. Honestly, the fanciest part is probably the CSS animations.

I looked at Ink and Twine early on but ended up rolling my own because I wanted tight control over the visual presentation and the stat tracking, plus it would’ve been one more thing to learn while my energy was high. Whether that was the right call or just a classic case of “I’ll build it myself, how hard can it be”, jury’s still out.

If you get a chance to play further, I’d love to hear what you think! The arc is six nights, maybe 3-4 minutes each.

6

This is really helpful, thank you for playing through it twice!

Regarding the AI detection: yeah, that’s a fair assessment. I’m not going to pretend that’s not there. The prose is AI-assisted (Claude, with a lot of editing and a craft review system I built to catch the worst tells). I did a pretty significant pass on the prose this morning actually, targeting the patterns that make it read as AI, like structural symmetry, over-polishing, that kind of thing. So your timing is good in that you probably played the version right before I fixed a bunch of what you’re describing. Whether the new version actually lands better… I hope so, but genuinely don’t know yet. I’d be curious if you ever felt like poking at it again, but no pressure at all; this is valuable feedback!

The choices feeling cosmetic is a harder one to hear, but I think you’re right. This is a 6 night arc, designed as a playable prototype, to validate there’s interest, an audience, and whether I could do it. The stats will move in the full game (if it gets enough validation to be built), and they do affect which prose you see and how the arc resolves. Right now, there are three distinct paths with different endings, but if you can’t feel that divergence while you’re playing, that’s on me, not on you for missing it. Something I need to make feel more apparent.

The faction lore is interesting because you’re the third person to flag it independently. Kron, Skein, the Coil; the intent was to introduce them gradually so you’re learning the world the way a new vampire would, but “gradually” might have tipped into “confusingly.” I did have an intro on the landing page, but removed it because it was a lot of scrolling before getting to the “play”. I think there’s a middle ground between info dump and mystery box that I’m grappling with.

Really appreciate you taking the time! This is exactly the kind of feedback that’s hard to get from friends who are trying to be nice.

7

Wow - This is a lot to chew on, thanks! And thank you for playing through it three times! It’s way more than I expected, and your detail here is genuinely useful.

Let me take these one at a time:

Regarding the AI prose: you’re right, and I’m not going to argue with that. It’s AI-assisted (Claude, plus a review pipeline I built to catch the worst machine patterns). I actually did a big revision pass this morning targeting exactly the stuff you’re flagging, like the structural tells, over-polish, and clichés. You played the pre-revision version. I’m not claiming the new version would change your mind, but if you have the time, another look would be super helpful.

The “results, buildup, two options” formula: this one stings - but I need to hear it - because you’ve identified the structure, and if it’s *that* visible, it’s not doing its job. The arc does have three genuinely different paths with different endings (Docks Clean, Meridian Composed, Rupture), but if the branching reads as picking an archetype rather than making a decision, that’s something structural I need to look into.

“Why is the climax called the climax”: ha ha. Fair. I honestly thought it’d be helpful for players to see the arc, to understand how things are structured, and to know there’s more coming. Let me look into reframing it.

The faction confusion: you’re (now) the fourth person to tell me the Kron/Skein/Coil stuff is too opaque and, to quote my husband: “not sharp enough”. At some point I have to stop calling it “intentional mystery” (for the sake of the prototype) and admit I haven’t given people enough to hold onto. I did have more intro material on the landing page, but removed it because it was a lot of scrolling before you even got to “play.” That created its own problem, where people were bouncing before they ever saw a choice. Still looking for the middle ground (I suppose that’s a never-ending battle). The woman you’ve “seen twice”: if that reference didn’t connect to anything for you, there’s a continuity issue I need to track down.

Slow animations and the italic body text: both noted, both fixable. You’re right that the italic reads as pullout quote.

Showing location per night when you can’t control where you go: I actually hadn’t thought about this from the player’s side, which is kind of embarrassing now that you say it. It’s there because the full game design has district choices that matters to the plot. In the prototype arc it’s less important.

For context on some of this: this is a 6-night prototype arc, built to validate whether there’s interest and an audience before I build the full game. So some of what feels constrained is genuinely pared down. But, I don’t want to use that as me dodging the feedback, because the choices feeling predictable isn’t a scope problem.

Anyway. Thank you for being this thorough! It is seriously impressive and appreciated!

1 Like
8

I took a look at a couple passages. It’s still clearly AI-generated (to me) but at least most of the main tells are gone.

1 Like
9

Ha! What’s still tipping you off? I’ve been working on the specific tells. Curious what’s left that I’m not catching

10

I can’t offer any specific tips here, but I suspect trying to hand-edit LLM-generated text to remove all the tells of being LLM-generated will end up being more work than just writing your own text.

4 Likes
11

Hello. I tried the three demos. I thought the writing was good and the choices relevant, but that they didn’t really go anywhere.

Presumably, these are meant to be a “taster” because, at the end of the nights, there isn’t any real conclusion. It felt, instead, the story had only just started.

Presentation is good.

I’m not really a stats player, so i mostly ignored them and just make my choices on whatever i felt was right. Whether or not the stats had actual consequences, i didn’t know. Nor did the game indicate (eg by greyed out choices).

At any time, there were only two choices. I guess this can work, but it started to feel rather limiting.

Looks good, good Luck with it.

1 Like
12

Yeah, if you’re so dead-set on removing the tells, maybe not write with AI in the first place. I could give you a line-by-line analysis of a passage but I really don’t see a reason to.

3 Likes
13

It starts with a wordy introduction, describing something using a lot of words, some of which might actually be necessary, but probably not as many words as are actually being used in this particular case right now. Then some short sentences. That aren’t even sentences. A clarification. Not this thing, a different thing. A phrase leading up to an em-dash — and the occasional oddly overwrought simile, like a Gothic cathedral in the middle of a cornfield.

(Granted, all this could also be a stylistic choice, but a decent number of AI generated stories that I’ve read have been styled like this, so that’s where I got the AI generated vibe)

6 Likes
14

and @Draconis - Totally fair points. Appreciate the honesty.

15

Thank you for playing all three! I really appreciate you giving it that much time.

The “story had only just started” feeling is intentional for the prototype, but I think I’m not setting that expectation well enough. Each night is meant to be one short chapter: a few minutes, one or two choices, then you come back tomorrow (or advance the story; figuring out mechanics) and see what followed. I’m leaning toward a serialized thing than a complete story in one sitting, but again, feedback like yours is super helpful in figuring out what folks want.

Can I ask?: when you landed on the page, was there anything that told you it was meant to be a short nightly session? I have a primer on the landing page, but I’m starting to think people are skipping past it, or it’s not clear enough. If it wasn’t visible to you, that’s useful to know.

The stats point is fair too. Right now there’s no indication that stats affect anything downstream, so ignoring them is the rational move. That’s on me to fix when/if there are enough signals to build the full game.

Two choices per night is a scope constraint for now, I totally get it. The full game would have more options, depending on where you are, and what you’ve done. I hear you that it starts to feel narrow after a few rounds.

16

Very helpful, thank you for breaking it down!

The em dash thing and the “not X, Y” construction are both on my list already. The overwrought simile pattern is something I’ve been watching for too. But the rhythm you’re describing: long wordy setup, I hadn’t thought of as a distinct pattern. That’s a good catch.

You’re right that some of it is also just… how I write? The fragment thing especially. But knowing that it reads as AI-generated to people who’ve seen a lot of AI prose is useful regardless; the effect on the reader is the same either way.

1 Like
17

I actually clicked the link from the top of the landing page without reading any more. This is what the page wants me to do. Also, it was a bit confusing which link to click, although it turns out they do the same thing.

1 Like
18

It does say “the last ten minutes of your night” and “play your first night”.

19

@robserpa I played your game, and while it took me a few passages to get the AI vibe, that realization didn’t sour me. What did disappoint me was the problems inherent with all creative AI content I’ve ever encountered, and that is: AI doesn’t have anything to say.

After playing a session, I felt that there was nothing rewarding that happened. There was nothing to ponder or challenge my assumptions. Nothing to beckon me back. Ultimately, the story said… nothing.

However, even if you manage to get AI to say something worthwhile… it still has to deliver it in a satisfying manner. The story felt like it had only one speed and one tone for the most part. Pacing is such an important aspect of writing, and it’s so nuanced. You can’t quantify it. You’ll only know it’s done well when you experience it.

I believe this is what @Hidnook and @Draconis were getting at. Forget about the stigma that AI carries and the superficial writing patterns it employs; just solve the actual problems AI brings to the quality and delivery of a story. If you do that, the story will be satisfying and readers will praise your game. If you focus on hiding the fact that it’s AI and the story still falls flat, what will you have accomplished?

Food for thought.

5 Likes
20

Howdy HAL - This is one of the most useful things anyone’s said in this thread, so thank you for taking the time.

You’re right about the pacing; One speed, one tone. The arc was built to test whether a nightly format creates pull, and I think I over-optimized at the expense of character/emotional range that’s appealing to my audience.

The “nothing to say” critique is harder for me to hear, though I absolutely need to process it. The current arc is six short chapters that test the mechanics and prototype.

I think where I’d push back slightly is on the framing that this is inherent to AI: The ideas and the emotional architecture are all mine, along with the pacing failures. If the story has nothing to say, that’s my fault, not the tool’s. And I hear the deeper point, that if the tool makes it easy to produce competent prose that doesn’t challenge anyone, it might be making it harder to notice when the story is empty. That’s a real risk that I’m going to think about.

Appreciate the directness.

2 Likes