Prerequisites
- WhereScape 3D version 9.0.6.0 or later (The version is enforced at runtime).
- Python 3.9 or later.
- An API key (or equivalent credentials) for your chosen LLM provider.
Python Packages Required
| Package | Description |
|---|---|
| langchain | The core LangChain framework for building LLM-powered applications. Provides base abstractions for chains, agents, memory, and tools. |
| langchain-openai | LangChain integration for OpenAI and Azure OpenAI models. |
| langchain-anthropic | LangChain integration for Anthropic's Claude models. |
| langchain-google-genai | LangChain integration for Google's Gemini models via the Generative AI API. |
| langchain-aws | LangChain integration for AWS Bedrock, providing access to multiple model families hosted on AWS. |
| langchain-mistralai | LangChain integration for Mistral AI's hosted models. |
| langchain-groq | LangChain integration for Groq's high-speed inference API. |
| langchain-xai | LangChain integration for xAI's Grok models. |
| langchain-ollama | LangChain integration for Ollama, enabling use of locally-run open-source models. |
| requests | A widely used HTTP library for making API calls and web requests in Python. |
| tqdm | Adds progress bars to loops and long-running tasks, useful for tracking processing progress. |
| tenacity | A retry library for Python, used to automatically retry failed operations with configurable wait strategies and stop conditions. |
| boto3 | The official AWS SDK for Python. Required by langchain-aws to authenticate and communicate with AWS Bedrock. |
Supported LLM Providers
| Provider | Description | Documentation |
|---|---|---|
OpenAI | The most widely used AI provider, offering the GPT model family. Industry standard with broad tooling and ecosystem support. | Available models: https://platform.openai.com/docs/models OpenAI Services Agreement, effective January 1, 2026: https://openai.com/policies/services-agreement OpenAI Service-Specific Terms: https://openai.com/policies/service-terms OpenAI Usage Policies: https://openai.com/policies/usage-policies OpenAI Data Processing Addendum: https://openai.com/policies/data-processing-addendum |
| Anthropic | Creator of the Claude model family, with a strong focus on AI safety and alignment. Excellent for long-context tasks. | Available models: https://docs.anthropic.com/en/docs/about-claude/models/overview Anthropic Commercial Terms: https://www.anthropic.com/legal/commercial-terms Anthropic Usage Policy: https://www.anthropic.com/legal/usage-policy Anthropic Supported Regions Policy: https://www.anthropic.com/supported-countries Anthropic Privacy and Data Processing materials: https://www.anthropic.com/legal/privacy |
| GoogleGenAI | Google's Gemini model family via the Generative AI API. Notable for its large context windows (up to 1M tokens). | Available models: https://ai.google.dev/gemini-api/docs/models Gemini API Additional Terms: https://ai.google.dev/gemini-api/terms Google APIs Terms of Service: https://developers.google.com/terms Google Generative AI Prohibited Use Policy: https://policies.google.com/terms/generative-ai/use-policy Google Privacy Policy: https://policies.google.com/privacy Google data treatment may differ based on service and account type. |
| AzureOpenAI | Microsoft's enterprise-grade deployment of OpenAI models via Azure. Offers enhanced security, compliance, and SLA guarantees. | Available models: https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models Microsoft Azure Legal Information: https://azure.microsoft.com/en-us/support/legal/ Microsoft Product Terms: Microsoft Products and Services Data Protection Addendum: https://www.microsoft.com/licensing/docs/view/Microsoft-Products-and-Services-Data-Protection-Addendum-DPA Microsoft Privacy Statement: https://privacy.microsoft.com/privacystatement Microsoft Enterprise AI Services Code of Conduct: https://aka.ms/AI-CoC |
| Bedrock | AWS's managed service providing access to multiple model families (Anthropic, Meta, Mistral and more) via a single API. Integrates natively with AWS infrastructure. | Available models: https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html AWS Customer Agreement: https://aws.amazon.com/agreement/ AWS Service Terms: https://aws.amazon.com/service-terms/ AWS Acceptable Use Policy: https://aws.amazon.com/aup/ AWS Privacy Notice: https://aws.amazon.com/privacy/ Amazon Bedrock Data Protection documentation: https://docs.aws.amazon.com/bedrock/latest/userguide/data-protection.html Amazon Bedrock may provide models from multiple model providers. Model-specific EULAs may be presented when the customer subscribes to a model. Customers using Bedrock remain responsible for reviewing both the AWS agreement and any model-provider terms applicable to the selected Bedrock model. |
| Mistral | A European AI provider offering efficient, cost-effective models. Strong multilingual capabilities and a good balance of performance vs price. | Available models: https://docs.mistral.ai/getting-started/models/models_overview Mistral legal terms portal identified in the submission: https://legal.mistral.ai/terms Mistral privacy policy: https://legal.mistral.ai/terms/privacy-policy |
| Groq | Known for extremely fast inference speeds using custom LPU (Language Processing Unit) hardware. Best choice when low latency is a priority. | Available models: https://console.groq.com/docs/models Groq Terms of Use: https://groq.com/terms-of-use Groq Privacy Policy: https://groq.com/privacy-policy Groq legal documentation: https://console.groq.com/docs/legal Use of the Groq API with WhereScape 3D Model Assist AI is solely subject to the customer’s Groq agreement, WhereScape has not validated the terms applicable to the customer’s account. |
| xAI | Elon Musk's AI company offering the Grok model family. Positioned as an alternative to GPT with real-time web access capabilities. | Available models: https://docs.x.ai/docs/models xAI Legal Portal: https://x.ai/legal xAI Privacy Policy: https://x.ai/legal/privacy-policy |
| Ollama | A local model runner that lets you run open-source models (Llama, Mistral, Gemma etc.) entirely on your own hardware. No API key needed, ideal for privacy-sensitive use cases. | Available models: https://ollama.com/library Ollama Terms of Service: https://ollama.com/terms Ollama Privacy Policy: https://ollama.com/privacy Ollama model library: https://ollama.com/library Support for Ollama does not include a license to any model. |
Support for specific AI providers and models is subject to change at any time.
Choosing an AI Provider
GPT-4o and Claude Sonnet 4 consistently produce the most reliable and well-structured JSON output. If you are evaluating providers, start with one of these before testing others.
Running The Installation Script
- Extract the '3D_Model_Assist_AI.zip' file after downloading it from the WhereScape support portal.
- Execute the 'Install_WS3D_EP_ddmmyyyy-hhmm.ps1' script, ensuring that it is run as administrator. If you are unable to run the script as an administrator you may need to install the Python dependencies listed above manually.
- Select each of the repositories that you wish to use in combination with 3D Model Assist AI.
- Review the logging details for progress updates.
- Once the script has completed successfully open WhereScape 3D and apply the final configurations.
Final Configuration
- Apply 'LLM connectivity' repository level UI configuration file on the repositories general settings tab. This can be found by navigating to Tools → Settings in 3D.
- Click 'Apply' on the settings dialogue.
- Apply the 'Data Vault Design - 3D Model Assist AI' version level workflow set in the Data Vault Design category. This can be found by navigating to Tools → Settings in 3D.
- Enter the information required to connect to your LLM provider.
- Add a documentation field named '3D Model Assist AI' at the column level in the data vault design settings.
