Ghelia Launches 'GHELIA AutoDeck,' a 3D Model-Based Knowledge Platform for Structuring and Streamlining Design Data via LLMs
Ghelia Inc. has released 'GHELIA AutoDeck,' a manufacturing platform that uses multimodal LLMs to interpret and store unstructured design data, such as CAE and 3D models. Operating on-premise to ensure security, it enables the codification of tacit knowledge to improve design efficiency.
📋 Article Processing Timeline
- 📰 Published: May 21, 2026 at 01:20
- 🔍 Collected: May 20, 2026 at 17:02
- 🤖 AI Analyzed: May 20, 2026 at 17:23 (21 min after Collected)
## Launch of GHELIA AutoDeck
Ghelia Inc. (Headquarters: Taito-ku, Tokyo; President and CEO: Makoto Saito) has announced the launch of 'GHELIA AutoDeck' (GAD), a platform designed to accumulate and manage unstructured design-related documents, simulation data, and image data using multimodal LLMs.
## Background
As the manufacturing environment undergoes significant changes, the loss of enterprise-specific knowledge in product development has become a critical issue. Ghelia aims to establish a knowledge foundation for design and development, allowing manufacturers to accumulate, inherit, and optimally utilize valuable knowledge.
## Key Features
GHELIA AutoDeck utilizes multimodal LLMs to ingest and generate descriptions for unstructured data such as 3D models, simulation results, and images. For CAE data, by processing simulation result images along with analytical parameters, boundary conditions, and logs, the platform automatically generates natural language descriptions covering the 'objective,' 'observations,' 'discussions,' 'conclusions,' and 'next actions.'
## Addressing Previous Limitations
Historically, accumulating design know-how was difficult due to the complexity of unstructured data, time-consuming documentation, and inconsistent reporting. Attempts to use LLMs or RAG on raw data often yielded low-accuracy results due to the inability to handle unstructured formats effectively. GHELIA AutoDeck addresses this by using multimodal LLMs during the ingestion phase to transform tacit knowledge into structured, actionable intelligence.
## Security and Future Outlook
Prioritizing data privacy, GHELIA AutoDeck operates in an on-premise environment. By accumulating design intent and insights from discarded designs, the platform supports the realization of a 'Digital Thread.' Ghelia will exhibit and demonstrate the product at the 'Manufacturing World Tokyo' event from July 1 to 3, 2026, at Tokyo Big Sight (Booth No. W10-31).
Ghelia Inc. (Headquarters: Taito-ku, Tokyo; President and CEO: Makoto Saito) has announced the launch of 'GHELIA AutoDeck' (GAD), a platform designed to accumulate and manage unstructured design-related documents, simulation data, and image data using multimodal LLMs.
## Background
As the manufacturing environment undergoes significant changes, the loss of enterprise-specific knowledge in product development has become a critical issue. Ghelia aims to establish a knowledge foundation for design and development, allowing manufacturers to accumulate, inherit, and optimally utilize valuable knowledge.
## Key Features
GHELIA AutoDeck utilizes multimodal LLMs to ingest and generate descriptions for unstructured data such as 3D models, simulation results, and images. For CAE data, by processing simulation result images along with analytical parameters, boundary conditions, and logs, the platform automatically generates natural language descriptions covering the 'objective,' 'observations,' 'discussions,' 'conclusions,' and 'next actions.'
## Addressing Previous Limitations
Historically, accumulating design know-how was difficult due to the complexity of unstructured data, time-consuming documentation, and inconsistent reporting. Attempts to use LLMs or RAG on raw data often yielded low-accuracy results due to the inability to handle unstructured formats effectively. GHELIA AutoDeck addresses this by using multimodal LLMs during the ingestion phase to transform tacit knowledge into structured, actionable intelligence.
## Security and Future Outlook
Prioritizing data privacy, GHELIA AutoDeck operates in an on-premise environment. By accumulating design intent and insights from discarded designs, the platform supports the realization of a 'Digital Thread.' Ghelia will exhibit and demonstrate the product at the 'Manufacturing World Tokyo' event from July 1 to 3, 2026, at Tokyo Big Sight (Booth No. W10-31).
FAQ
What is the purpose of GHELIA AutoDeck?
To digitize and codify manufacturing design expertise that often becomes siloed or lost, enabling efficient knowledge transfer and utilization.
What are the benefits of on-premise operation?
It ensures the confidentiality of proprietary design data and technical information while providing a secure, dedicated knowledge foundation.
Does it handle more than CAE data?
Yes, it can process multimodal formats including 3D models, image data, and various design-related documents.