AI Frameworks · AIScientists-Dev

WorldSeed

WorldSeed is a Python-based multi-agent simulation engine where AI agents interact within a declaratively defined world, producing emergent behaviors. It enables building everything from research simulations to game worlds without hardcoded domain logic.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/AIScientists-Dev/WorldSeed

861

GitHub stars

Forks

Python

Primary language

MIT

License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	AIScientists-Dev/WorldSeed
Owner	AIScientists-Dev
Primary language	Python
License	MIT — OSI-approved
Stars	861
Forks	55
Open issues	0
Latest release	Unknown
Last updated	2026-05-08
Source	https://github.com/AIScientists-Dev/WorldSeed

What WorldSeed is

A tick-loop agent framework with YAML-defined world state, asymmetric perception filters, deterministic rules via DSL, and LLM-backed decision making ('Dungeon Master') for uncertain outcomes. Supports custom agents via OpenClaw or Codex subagents; accepts any LiteLLM-compatible model.

Quickstart

Get the WorldSeed source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/AIScientists-Dev/WorldSeed.git2cd WorldSeed3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Multi-agent research automation

Autonomous research teams that formulate hypotheses, run experiments, peer-review, and cite each other. Useful for exploring hyperparameter spaces, architecture variants, or structured exploration problems where emergence matters.

Interactive narrative & game worlds

Real-time multiplayer or single-player worlds with asymmetric information, hidden agendas, and player intervention. Agents respond to rules and LLM judgment; humans can observe, whisper privately, or play as a character.

Organizational scenario simulation

Simulate workplace dynamics, resource allocation, conflict resolution, or policy impacts at organizational scale. Each agent has private knowledge; outcomes blend deterministic rules and AI judgment.

Implementation considerations

Requires fluency in YAML-based world declaration; complex scenarios need hand-crafted YAML or AI-generated scaffolding (create-world skill). Plan for iterative world tuning.
Agent implementation: choose OpenClaw or Codex subagents. Documentation references both; verify which fits your agent requirements and LLM provider.
LLM cost and latency: every uncertain action requires a DM call. Budget for token consumption and consider fallback logic if API is unavailable.
Perception filters are per-character and declared in YAML; asymmetric information requires careful schema design to avoid leakage or missed nuance.
Replay and introspection: all state changes logged. Plan storage for multi-hour runs; past runs are replayable but persistence strategy not detailed.

When to avoid it — and what to weigh

Requiring real-time performance at scale — Each uncertain outcome requires an LLM call (network latency). No release notes or benchmarks provided on throughput or latency. Large agent counts may bottleneck on model inference.
Need deterministic reproducibility without LLM variance — LLM-based 'Dungeon Master' introduces non-determinism for uncertain outcomes. If you need bit-for-bit reproducibility or minimal model dependency, this adds complexity.
Seeking off-the-shelf SaaS integration — Self-hosted only (no mention of managed service). Requires Python 3.11+, Node.js 18+, uv, and an external LLM API key. Not a plug-and-play SaaS product.
Production systems with formal audit/security requirements — Project is ~3 weeks old (created 2026-04-15). No security audit, compliance documentation, or formal SLA history. Early-stage for regulated use.

License & commercial use

MIT License. Permissive OSI license; no restrictions on commercial use, modification, or distribution. Include license text in derivative works.

MIT is a permissive license allowing commercial use without royalties or attribution requirements. No mention of trademarks, trademark use guidelines, or commercial support terms. If building a product, review trademark policy with the maintainers (morphmind.ai) independently.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Adequate
License clarity	Clear
Deployment complexity	Moderate
DEV.co fit	Good
Assessment confidence	Medium

Security considerations

Project age (3 weeks) means security review is limited. Key considerations: (1) LLM prompt injection via agent perception or DM queries—ensure input validation and sandboxing if agents can manipulate their own perception filters. (2) API key exposure: .env.example shown; ensure secrets are not logged. (3) No mention of authentication, authorization, or multi-tenant isolation. (4) Agent sandbox: no mention of code execution limits or resource quotas if agents run custom code. (5) Audit trail: logged for replay, but no tamper-evident or immutable log design described. Suitable for non-regulated research/games; avoid for sensitive PII or compliance-critical work without additional hardening.

Alternatives to consider

Mesa (Python multi-agent simulation framework)

Mature Python framework for agent-based modeling. Stronger on spatial simulation, statistics, and educational use. Less opinionated on narrative/game logic; no built-in LLM integration.

Langchain Agents + LangGraph

Lower-level primitives for chaining LLM calls and managing agent state graphs. More flexible for custom logic; requires more scaffolding. No world-state engine or YAML config.

Custom multi-agent system (OpenAI Swarm, Anthropic's tools API + orchestration)

Direct use of LLM APIs with custom orchestration. Full control; no framework overhead. Requires building world simulation, perception, and tick loop yourself.

Software development agency

Build on WorldSeed with DEV.co software developers

Try the interactive demo, review the YAML-based world definition, and assess fit against your simulation or game goals. Consider early-stage maturity and LLM cost before production adoption.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

WorldX WorldGen dynamiq

Related on DEV.co

Explore the category and the services that help you build with it.

AI Frameworks AI Application Development Custom Software Development API Development Talk to DEV.co

WorldSeed FAQ

Can I run this without an LLM API?

Partially. Deterministic outcomes (rules-based) resolve via DSL with no LLM call. Uncertain outcomes require a DM call. You could mock the DM with a simple rule engine, but that defeats the emergent-behavior value proposition.

What LLMs are supported?

Any LiteLLM-compatible provider: OpenAI, Anthropic, Ollama, Cohere, etc. LiteLLM is a wrapper; support scope depends on LiteLLM's latest version. Offline/local models possible via Ollama.

How much does it cost to run a simulation?

Unknown without benchmarks. Cost depends on: number of agents, number of ticks, complexity of uncertain outcomes, and LLM pricing. A real example (Autoresearch: 100 hypotheses, 86 experiments in 11 hours) is in the README; cost not disclosed.

Can I deploy this to production?

Technically yes (it's code). Practically, plan for: no formal SLA history, 3-week age, no horizontal scaling docs, no auth/multi-tenancy, no security audit. Suitable for research/beta; risky for regulated or revenue-critical use without hardening.

Custom software development services

Need help beyond evaluating WorldSeed? DEV.co is a software development agency offering software development services and web development for teams of every size. Our software developers and web developers build custom software, web applications, APIs, and ai frameworks integrations — and maintain them long-term.

Explore WorldSeed for Your Next Multi-Agent Project

Try the interactive demo, review the YAML-based world definition, and assess fit against your simulation or game goals. Consider early-stage maturity and LLM cost before production adoption.

Talk to DEV.co Schedule a Consultation