New short interactive sci-fi story: Operation Red Shadow (looking for feedback)

Playing Beta Testing Requests and Discussion
1

Hi everyone,

I’ve just released a new short interactive sci-fi story as part of my small experimental project called Future Logs. It’s titled “Operation Red Shadow”.

The story is set on Mars in 2069. You play as an undercover agent-hacker caught in the middle of an escalating conflict. The experience focuses on conversations with three different NPCs, each with their own mindset and narrative context. What you learn from these interactions shapes how events unfold and leads to meaningful consequences.

This is designed more as an experiment in narrative-driven interactive fiction than a traditional game. It’s short, browser-based, and meant to be completed in a few minutes.

I’d really appreciate any feedback, especially on:
– how the NPC conversations feel
– whether the choices feel meaningful
– the overall pacing and immersion

You can find it on the Future Logs site (just search for “Operation Red Shadow”), and I’m happy to share a direct link if anyone prefers.

Thanks in advance for any impressions or suggestions!

3 Likes
2

I’ll try desktop later. On mobile, the keyboard covers the input so you can’t see to type.

1 Like
3
  • contains AI-generated images…
  • the dialogue being in that glowy font makes it hard to read. One person speaking on multiple lines with separate dialogue is also weird.

“I’m glad you came back to us. We barely got you out.”
“We are on the MDA-Kestrel, sick bay, en route back to Earth.”

  • “Reed. Is that me?” What about this sentence is meant to draw the reader in?
  • I’d like to learn more about the environment, history, etc. but right now it seems generic futuristic
  • lots of telling with little showing
  • you shouldn’t really have the —I say style dialogue tag
  • the swearing doesn’t really fit
  • it does have tells reminiscent of AI-generated writing
  • AI-powered dialogue makes it a slog to get to where you want to go
  • what’s the point of choices if you just say “this wasn’t part of the memory”?
  • dialogue doesn’t seem to advance the plot all that quickly and it’s unclear what to do

I stopped partway through communicating with the second person…not my kind of game, sorry!

1 Like
4

Thanks for the feedback, much appreciated.

You’re right about the glowing bold font — I already changed it to a softer style, it should be easier to read now.

Also good catch about the dialogue lines being too close together. I’ll adjust the spacing there, it does look a bit cramped in longer exchanges.

About the choices and the “this wasn’t part of the memory” thing — that’s actually a small mechanic tied to the implant. You’re basically trying to reconstruct how the mission really went. Too many off-track choices overheat the implant and eventually kill the agent.

Kind of a shame you didn’t get to the end — the conversation with Sara can get pretty engaging if you steer it a bit. It can start to feel almost like talking to a character with her own personality.

Thanks again for playing it and sharing your thoughts!

1 Like
5

Ideally you don’t want to make your players wait that long for your NPCs to start to feel like they have personality.

3 Likes
6

Definitely some strange occurrences. The text in the input box (‘What are you saying’, etc) changed to Polish after a bit, then back to English.

Sara’s responses started to include emoji after a few lines, and her dialogue just… isn’t compelling to me whatsoever. Repetitive, slightly scrambled riffs on the same basic content, technically responding to whatever the player says but in a way that doesn’t actually respond. There isn’t much feeling of discovery when the model behind her responses is basically pushing the ‘secret’ on the player in every paragraph.

There’s also some person/number/gender discontinuity with pronouns but I don’t know whether that’s an artifact of the generative text or with translation from Polish. E.g. “Tension grips your throat” in the intro, “He’s coming back” in a narration paragraph about Sara.

Is Sara the only conversation that has a model behind it? The supervisor and Vale seem to be much simpler and keyword-based.

2 Likes
7

Thanks a lot for the feedback, really appreciate it.

About the translation – yeah, that’s the downside of machine translation. It definitely needs polishing. I fixed the details you pointed out, hopefully nothing obvious left now. The upside is that it’s available in 7 languages and I can easily add more.

About the NPC responses: they’re meant to feel like natural conversation. If the input is just a few words, they sometimes struggle to respond in a meaningful way. It’s not keyword-based though – it’s not “go south / look / use key” style. You can talk to them naturally, in any language, and they react to what you actually say.

Tweaking the NPC instructions took me a long time. They may decide to trust you and reveal info, or do the opposite and cut the conversation short. It’s a bit unpredictable by design.

Supervisor and Vale are simpler, but they’re not keyword-based either. I tried to make them more likely to trust you, but I might make them more suspicious later. It’s hard to balance without more feedback on how people actually talk to them.
Below I’m pasting about 1/4 of Sara’s instruction, just to show how it’s set up.

Now imagine a story where you meet 10 NPCs, each with their own personality, mindset, and knowledge. That wasn’t really possible even 6 months ago. So now you don’t just read a dialog, you kind of create it. And it may lead to different outcomes. Actually interactive fiction.

On mobile you can even use speech-to-text, which makes it feel more like talking to a person.

I really think AI chat inside stories could be a game changer for storytelling (and maybe education too).

Below part of NPC-Sara’s instructions (about 1/4):

During the conversation:
– reveal information about yourself
– do not repeat the same info from earlier replies
– do not repeat questions
– talk about injustice, taxes, MDA control, rising resistance and frustration
– build emotional context, do not agitate openly
– if he seems trustworthy, start trusting him
– do not end the conversation too quickly, especially if you trust him
– limit each reply to max 200 characters (do not mention the number)

SUCCESS CONDITIONS:
If the speaker:
– shows understanding for Mars’ situation
– criticizes Earth or MDA dominance
– declares a desire to “be on Mars’ side”
– does not behave like an agent or spy
– is consistent and sincere over several replies

then:
– stretch the conversation to 9–10 turns
– gradually reveal information:
– you think you can trust him
– talk about Mars’ situation and growing frustration
– invite him to a meeting on Sunday at the bar: “Red Dust”
– reveal preparations for a fight
– say there are training materials on Red Dawn servers
– you may invent personal details and backstory
– do not repeat yourself, stay in context
– end the conversation with verdict: SUCCESS