Living Image - limg

What the company appears to be: Living Image (limg) is a self-reported prototype project that turns near-frontal anime portraits into portable, code-controlled character assets (.limg files). These assets can be loaded in a browser-based viewer and controlled via an API without requiring further machine learning inference at runtime. The system includes a compiler (Python-based), a .limg file format, and a browser-based runtime (TypeScript/Canvas).

What changed: The author reports having built this as part of a hackathon submission. It is described as a prototype with no revenue or customer traction. The project was submitted to the OpenAI 2026 hackathon.

Single most important open question: Is there evidence that the .limg format will be adopted by other developers or applications beyond this prototype?

The description states that Living Image is a system that:

• Takes an anime portrait (PNG/JPEG) as input.
• Uses a Python-based compiler to analyze facial landmarks and generate deformation data.
• Produces a self-contained .limg file containing:
  • Source texture
  • Geometry
  • Behavior parameters
  • Capabilities
  • Provenance information
• A browser-based viewer loads the .limg file and allows control through a normalized API (e.g., player.setState({...})).
• The viewer performs no ML inference at runtime; playback is deterministic and model-free after compilation.

Inference: The system separates model-backed compilation from deterministic playback, which implies a design choice to make the output portable and reusable across different platforms or applications.

The author claims that:

• Living Image enables "a favorite illustration" to be compiled once into a portable character asset.
• It supports programmatic control through an API.
• The system avoids traditional layered PSD workflows or manual rigging.
• Instead of generating one-off videos, it creates reusable assets.

Inference: The positioning appears to be that this is a tool for creating lightweight, reusable animated characters from static images — particularly suited for developers or creators who want to integrate such characters into apps or AI companions.

There is no evidence of prior market positioning or product evolution beyond the hackathon submission.

The description does not state specific customer segments or personas. However, it implies:

• Developers working on AI companions or interactive web experiences.
• Creators who want to animate illustrations programmatically.
• Users interested in integrating animated characters into games, apps, or digital assistants.

Inference: The target is likely technical users or developers who are building applications that require lightweight, reusable animated assets. No evidence of actual customers or user feedback exists.

There is no mention of pricing, monetization, or business model in the description.

Not evidenced: No indication of how this would be sold, licensed, or used commercially.

The project uses:

• Python for face detection and landmark analysis.
• PyTorch for ML inference during compilation.
• TypeScript/Canvas for runtime viewer.
• Docker and Cloudflare Workers for deployment.
• GPT-5.6 for architecture evaluation and documentation.
• Codex for research, implementation, testing, and packaging.

Inference: The technical stack suggests a hybrid approach combining ML-based preprocessing with deterministic runtime behavior. The separation of compilation from playback is intentional and documented.

The project is described as:

• A hackathon submission.
• A validated prototype.
• Not yet adopted by other applications or developers.
• No revenue, customers, or traction data provided.

Not evidenced: No evidence of adoption, usage metrics, or product maturity beyond the prototype stage.

No mention of competitors or existing solutions in the description. The author does not reference similar tools or platforms for animated character creation or animation asset management.

Not evidenced: No competitive landscape or differentiation analysis is provided.

• Prototype-only status: The system is described as a prototype, with no evidence of adoption or production use.
• No revenue or customers: There is no indication that the project has generated any income or attracted users.
• Unproven format adoption: The .limg format is not yet adopted by other developers or applications.
• Limited scope: The current implementation only supports near-frontal anime portraits and lacks broader art style support.
• No scalability claims: No evidence of plans to scale beyond the prototype or integrate with larger ecosystems.

1. What is the expected adoption rate for the .limg format among developers or creators?
2. Are there any early adopters or partners interested in integrating this technology into their products?
3. How does the project plan to expand beyond anime portraits and support other art styles?
4. Is there a roadmap for open-sourcing or standardizing the .limg file format?
5. What are the technical limitations of the current model, and how will they be addressed?
6. Has the team considered licensing or monetization models for the technology?