AIScientists-Dev

AIScientists-Dev / WorldSeed

Public

A world engine where AI agents live autonomously — define any scenario in YAML, watch stories emerge

worldseed.morphmind.aidemo agent-based-simulation agentic-ai agents ai ai-agents
52
2
100% credibility
Found Apr 17, 2026 at 52 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

WorldSeed is a simulation engine that brings user-defined worlds to life with autonomous AI agents, allowing observation, intervention, and custom scenario creation via a web dashboard.

How It Works

1
🔍 Discover AI worlds

Stumble upon WorldSeed demos where AI characters live out dramatic stories like office intrigue or spy games on their own.

2
🚀 Launch a ready world

Pick an example scenario and press play to bring a living office or teahouse to life with thinking agents.

3
👀 Watch stories unfold

Peer into the map, read agents' private thoughts, and see decisions ripple across the world in real time.

4
💬 Whisper or join in

Nudge an agent privately or step into a role to steer the drama without breaking the flow.

5
✏️ Build your own world

Describe characters, places, and rules simply, then watch new tales emerge every run.

🌟 Endless unique stories

Every playthrough creates fresh, unpredictable narratives you can replay or share forever.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 52 to 52 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is WorldSeed?

WorldSeed is a Python-based world engine that lets you define any scenario—characters, rules, perceptions, actions—in YAML, then unleashes AI agents to live and act autonomously in tick-based simulations. Stories emerge unpredictably as agents perceive asymmetric info, follow deterministic rules, or defer to an LLM "Dungeon Master" for edge cases, all viewable via a React dashboard at localhost:8000 with map views, event streams, and replay controls. Run it via CLI like `uv run worldseed play ai_layoffs.yaml` for instant worlds like office layoffs or spy teahouses.

Why is it gaining traction?

It stands out with zero domain assumptions—pure YAML configs make any world (no hardcoded maps or physics), plus god-mode interventions like whispering to agents or stepping in as a character. The isometric map, live tick history, and gazette generator turn raw sims into watchable narratives, beating rigid agent frameworks. Demos hook devs fast, showing emergent drama without setup hassle.

Who should use this?

AI researchers prototyping multi-agent behaviors, game devs testing procedural stories, or indie creators simulating social dynamics like GitHub world edit scenarios or world seed banks. Perfect for educators modeling economies or conflicts, or PMs visualizing team stresses in AI layoffs sims.

Verdict

Grab it for experiments—solid docs, pytest coverage, and MIT license make tinkering easy, despite 52 stars and 1.0% credibility signaling early days. Scale up once agent APIs stabilize.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.