Skip to content
AI Tool

world-model-mcp Review

world-model-mcp is a deterministic, persistent 3D/2D spatial world model for AI agents with entity tracking, object permanence, movement simulation, and expected view frustum projection.

shipped Sep 30, 2026freemium
world-model-mcp — product screenshot

Why it matters

1Offers a free tier for its open-source components.
2Provides a persistent 3D/2D coordinate store and topological graphs for spatial relations.
3Includes movement collision simulation and goal alignment bridge for task management.
4Features interactive simulation and visualizer tools.

About world-model-mcp

Platforms
Web, Node.js
Target Audience
Developers and researchers working on AI and spatial modeling.

Pricing Plans

Open Source
Free
  • • Locally hosted
  • • No telemetry
  • • Open-source under MIT License

Cost Examples

  • • API fees billed directly by providers for third-party models such as OpenAI or Google Gemini

Leadership

Not Listed
API DocsGitHubOpen Source

Specs

API Available

Yes, public API

overview

What is world-model-mcp?

world-model-mcp is a spatial world model tool that enables AI agents, developers, and researchers to manage deterministic, persistent 3D/2D environments. It provides capabilities for entity tracking, object permanence, movement simulation, and expected view frustum projection, facilitating advanced AI agent development and research.

features

Key Features of world-model-mcp

world-model-mcp provides a comprehensive set of features designed for building and simulating AI agent environments, focusing on spatial awareness and interaction.

  • Deterministic, persistent 3D/2D spatial world model for consistent simulations.
  • Entity tracking to monitor and manage individual agents and objects within the environment.
  • Object permanence, allowing agents to understand the continued existence of objects even when not directly observed.
  • Movement simulation with collision detection for realistic agent interactions.
  • Expected view frustum projection to simulate what an AI agent can perceive.
  • Persistent 3D/2D coordinate store for reliable spatial data management.
  • Topological graphs for representing and analyzing spatial relations.
  • Goal alignment bridge for integrating task management with the world model.
  • Interactive simulation and visualizer tools for development and debugging.

use cases

Who Should Use world-model-mcp?

world-model-mcp is designed for individuals and teams engaged in the development and research of AI systems that require robust spatial understanding and interaction capabilities.

  • AI agents requiring a deterministic and persistent spatial world model for navigation and interaction.
  • Developers working on AI and spatial modeling applications that need entity tracking and object permanence.
  • Researchers in AI and spatial modeling seeking tools for movement simulation and view frustum projection.
  • Teams building interactive simulations and visualizers for AI agent behavior.

how to use

How to Use world-model-mcp

To begin using world-model-mcp, developers and researchers can access its open-source components and integrate them into their AI projects. The platform supports Web and Node.js environments.

  • 1Access the open-source components via the GitHub repository at https://github.com/putervision/world-model-mcp.
  • 2Refer to the API documentation at https://putervision.github.io/world-model-mcp/ for detailed integration instructions.
  • 3Utilize the persistent 3D/2D coordinate store to define and manage spatial environments.
  • 4Implement entity tracking and object permanence for AI agents within the simulated world.
  • 5Configure movement simulation and expected view frustum projection for agent perception.
  • 6Integrate with Playwright for automation or SQLite for data persistence as needed.

pricing

world-model-mcp Pricing & Plans

world-model-mcp operates on a freemium model, offering its core components as open-source. While the software itself is free, users may incur API fees for third-party models if integrated.

  • Open Source: Free (includes core features and access to the GitHub repository).

Enjoying this? Get one like it in your inbox each morning.

one email a day · unsubscribe in two clicks · no third-party tracking

Pros

  • +Provides a deterministic and persistent 3D/2D spatial world model.
  • +Includes built-in entity tracking and object permanence for AI agents.
  • +Offers movement simulation with collision detection.
  • +Features expected view frustum projection for agent perception.
  • +Integrates with Playwright and SQLite for enhanced functionality.
  • +Available as open-source with a free tier.

Cons

  • −May require custom development for highly specialized physics or sensor models beyond its core offerings.
  • −Integration with third-party AI models may incur additional API costs.
  • −Primarily focused on abstract AI agents, potentially less suited for highly detailed physical robot simulations compared to specialized robotics simulators.
  • −Requires familiarity with spatial modeling concepts for optimal utilization.

Similar Tools

world-model-mcp vs Competitors

world-model-mcp differentiates itself from general-purpose engines and specialized simulators by focusing specifically on a deterministic, persistent spatial world model for AI agents, offering pre-built features like entity tracking and view frustum projection.

1

A complete, open-source game engine that provides tools for 2D and 3D game development, including physics, rendering, and scripting, suitable for building custom AI agent environments.

Godot offers a full game development environment, which is more general-purpose than world-model-mcp's specialized focus on AI agent world models. You gain flexibility and a large community, but might need to implement some AI-specific features like explicit frustum projection or advanced entity tracking yourself, rather than having them pre-built.

2

A powerful open-source robotics simulator designed for complex indoor and outdoor environments, offering robust physics, sensor simulation (including cameras), and support for various robot models.

Gazebo is highly specialized for robotics simulation, providing advanced sensor models and physics that might exceed world-model-mcp's scope in some areas. However, it might have a steeper learning curve and be more focused on physical robots than abstract AI agents, potentially requiring more effort to integrate with non-robot-specific AI frameworks.

3

A physics engine known for its speed and accuracy in simulating complex articulated bodies with contacts, widely used in robotics, biomechanics, and reinforcement learning research.

MuJoCo excels at high-fidelity physics simulation, which is a core component of world-model-mcp's movement simulation. It provides a robust foundation for deterministic interactions but requires more custom development to build out the full 'spatial world model' with entity tracking and view frustum projection beyond basic camera views, which world-model-mcp aims to provide out-of-the-box.

4
PyBullet↗

A Python module for physics simulation, robotics, and machine learning, offering a user-friendly API to create and interact with 3D environments and agents.

PyBullet provides a Python-native way to create 3D environments and simulate physics, making it accessible for AI researchers. Similar to MuJoCo, it offers strong physics and basic rendering but requires more manual implementation for advanced features like object permanence and sophisticated view frustum projection compared to world-model-mcp's more integrated approach.

More on Stork

Related AI Tools

Other tools in this category, matched by shared tags

One short daily email of tools worth shipping. No drip funnel.

one email a day · unsubscribe in two clicks · no third-party tracking

For builders

This page is doing a job for someone else’s tool.

AI agents read it. Buyers land on it. It answers in eight languages and over MCP. Your tool can have one like it — live in 24 hours.