framersai

Agent swarm simulation with emergent tool forging, HEXACO personality evolution, and deterministic kernels. Built on AgentOS.

23
1
100% credibility
Found Apr 18, 2026 at 23 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Paracosm is an open-source simulation engine that lets users define custom worlds and pit AI leaders with distinct personalities against identical challenges to observe emergent divergences.

How It Works

1
🔍 Discover Paracosm

You stumble upon Paracosm, a fun way to simulate worlds where AI leaders with different personalities shape the same starting place into wildly different futures.

2
🚀 Open the dashboard

Head to the live demo or launch the dashboard on your computer to start playing with ready-made worlds like Mars colonies or lunar outposts.

3
🌍 Pick your world

Choose a scenario like a Mars settlement, tweak departments like medical or engineering, and set basics like population and years to simulate.

4
👥 Create rival leaders

Design two commanders with clashing personalities—one bold innovator, one careful planner—and watch how their traits drive every choice.

5
▶️ Hit play and watch

Click run to launch parallel simulations; crises hit, departments advise and invent tools, leaders decide, and colonies diverge turn by turn.

📊 Explore the outcomes

Dive into reports, chat with virtual crew members, visualize family trees and moods, and see exactly why one leader thrived while the other struggled.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 23 to 23 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 paracosm?

Paracosm simulates agent swarms in closed-world scenarios like Mars colonies or lunar outposts, where AI leaders with HEXACO personalities make top-down decisions amid crises. Define worlds in JSON, compile via CLI with `npm run compile`, then run parallel sims via `runSimulation` API or dashboard to watch identical seeds diverge based on leader traits. Built in TypeScript on AgentOS, it handles event generation, department analysis, emergent tool forging, and deterministic state progression.

Why is it gaining traction?

This agent swarm framework stands out for runtime tool forging—departments build calculators judged by LLMs—plus personality drift, yielding replayable divergences without scripting every outcome. As an agent swarm open source GitHub repo, it hooks agent swarm AI tinkerers with live dashboard viz, SSE streaming, and cheap per-run costs (~$1-7). Pairs well with agent GitHub Claude code for custom hooks or agent swarm Claude coding flows.

Who should use this?

AI researchers prototyping agent swarms, game devs generating emergent NPC societies, strategy teams modeling CEO or general decisions under stress, or defense analysts running wargames. Ideal for agent swarm GitHub users bypassing agent GitHub Copilot CLI limits toward full agent swarm n8n-style orchestration.

Verdict

Grab it for agent swarm experiments—dashboard demos paracosm absolutely via side-by-side runs—but 23 stars and 1.0% credibility score flag early maturity: solid NPM/docs/live sim, sparse tests. Worth a spin if agent swarms intrigue you.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.