procedural generation – Telegram
procedural generation
121 subscribers
4.92K photos
1.44K videos
2 files
10.4K links
Created by @r_channels
Download Telegram
Deriving NPC reactions from formation instead of rolling for them

Working on a character generation system where behaviour isn't randomized. The pipeline:

Formation (key experiences) → Values (what they protect) → Properties (anchor, limits, defenses) → Reactions

The idea: if you know what shaped someone and what they'll protect, their response to pressure isn't random. It's constrained.

Same identity + same situation = same response type. The generation happens upstream in the formation. Everything else is derivation.

Built 30 characters to test it. Curious whether this counts as procedural generation or something else entirely.

I'll post the link in the comments.

https://redd.it/1pses1c
@proceduralgeneration
Top-down shots from my procedural map generator for my roguelike.
Each image is a different seed. The system is almost finished, but I’d love feedback on layout clarity, variation, or any small tweaks I could make
https://www.reddit.com/gallery/1ps5hfb

https://redd.it/1psh70g
@proceduralgeneration
Space filling curve with loops (Splined)
https://redd.it/1pt2hco
@proceduralgeneration
What’s your setup for using multiple AI models?

There are amazing AI models out there, but using them across different platforms felt messy and inefficient.
That’s why we created ALLGPT — one workspace to access multiple models without jumping between tools.

Would love to hear how others here handle this.

https://redd.it/1pvb633
@proceduralgeneration
I'd like to generate procedural terrain that's very realistic any sources?

This is for a planet. What's I've been doing so far has seemed so fake and unrealistic

https://redd.it/1pvx2jr
@proceduralgeneration
My attempt at procedural object placement using poisson disk sampling + noise (for filtering)

A few years back I ran into the problem of procedurally placing objects of varying sizes with no overlap. What worked well for me was Poisson disk sampling, combined with noise-based filtering (less messy than pure random, less rigid than grids).

I wrote up a short explanation and shared the code in a dev blog post (link’s on my game’s Steam page) in case it’s useful to anyone tackling something similar: https://store.steampowered.com/news/app/2223480/view/509600244355304032?l=english

https://preview.redd.it/hdexxgv3rk9g1.png?width=1848&format=png&auto=webp&s=708368126dae1ae2342a85049f26e7076574c1e3

https://preview.redd.it/7gbjkg96rk9g1.png?width=764&format=png&auto=webp&s=71bfd656a26cdf51ec877fbd845f113c33811474

https://redd.it/1pw86oy
@proceduralgeneration