I’m testing a small project called Diaries from the Future – a set of short, experimental interactive sci-fi stories. Right now there are 4 pieces online (a fifth one is in progress).
The stories are not linear. Some of them branch quite a lot (one has around 8 different paths, another about 5), but I try to keep the structure tight and focused rather than “open-world”. The format sits somewhere between micro-fiction and very light narrative games.
In the latest piece, “Neon Case”, I’m specifically testing a narrative device: instead of just reading or picking dialogue options, you interrogate an in-story AI system (a real chatbot interaction) to access an android’s memory. What you extract doesn’t completely rewrite the plot, but it unlocks the next part of the story and shapes how you understand what happened. I’d especially like feedback on whether this feels immersive or just gimmicky.
What I’m trying to learn:
– Does this kind of interaction add to the narrative, or break the flow?
– Do a few meaningful branches work better than lots of small choices?
– Does this still feel like interactive fiction, or something in-between?
The project is free and web-based: future-logs dot com
I’m not trying to market it – I’m genuinely looking for honest feedback, especially on “Neon Case” and the AI-as-narrative-element idea. Even short or critical comments are very welcome.
I played the “Neon Case” game. Regarding the “chatbot” interaction;
This was done quite well, but the responses are too slow. Some, I had to wait for ages before a reply. However, I did like the free-form questioning and the answers mostly worked well. Although some were a little vague, which i guess is a thing with “chatbot” design.
It did feel a bit of a change of gear. First, i’m in a choice-based narrative and then i’m suddenly typing stuff in. I played on a desktop where typing is easy, but this could be annoying for mobile users. Sorry, i didn’t try mobile. Nevertheless, i don’t think it doesn’t work. And it did add an extra dimension.
This is of course the problem with detective games when choice-based. How to not tell the player the answer?
I don’t know answers, but what i have done with mixing text input with choices is to limit the text input to a single word or short phrase rather than a sentence. For example, you might need to know a “name” or a “password” to get into a location or scene. Perhaps a number, eg for a phone or address.
Other general points;
I really like the game and the artwork. Out of curiosity, i have a similar game idea, working title “Flasher”, which is a dystopian future with a black market for stealing people’s minds and “flashing” them into others. I’m reminded by a picture that is coincidentally similar to one of yours.
Thank you for taking the time to try it and for the thoughtful feedback.
You’re absolutely right about the response time — that’s currently one of the biggest drawbacks. I’m using grok-3-mini with instructions to answer quickly and without long analysis, but there’s still an unavoidable delay. Unfortunately that part isn’t something I can really shorten at the moment.
I’m really glad you felt the immersion, though. My intention was very much to make it feel like an interrogation / conversation, not just “typing commands into a system”. On mobile, for example, it actually works quite nicely with speech-to-text, which pushes it even more toward that interview-like experience.
I’m also already working on the next story. This one will involve conversations with three different NPCs, each with a distinct style and role in the narrative. I should have a test version ready by the end of the week.
Thanks again for the feedback — it’s genuinely helpful.
I’ve switched the AI engine to grok-4-fast, and the chatbot responses are now almost instant. That removes most of the waiting time and makes the interrogation feel much smoother and less stressful.
If anyone had issues with pacing before, I’d really appreciate it if you could try it again and let me know how it feels now.
