|
Selecting Your AI SystemImagix 4D supports three alternative modes of integrating with an AI engine. External Cloud Services offers seamless high-speed cloud integration using API keys. Local AI Server provides maximum data security via a self-hosted, secure infrastructure typically utilizing dedicated hardware. Interactive Chat-Based is a zero-cost manual approach through a browser-based copy-paste workflow.Note that running AI Analysis in either External Cloud Services mode or Local AI Server mode under Windows requires Windows 10 or higher. Linux installations require versions equivalent to Ubuntu 22 or higher, or AlmaLinux 9 or higher. External Cloud Services (OpenAI & Google Gemini)This mode interfaces directly with well-known cloud providers via secure HTTP connections, utilizing models specifically optimized for source code and architectural analysis. When a user invokes an analysis, Imagix 4D packages relevant information about the target of the analysis, and sends it straight to the cloud API endpoint. The primary user interface element is the AI Analysis window, which displays a "Query going to the system" screen before immediately populating the structured markdown response once it arrives (typically within 10 to 20 seconds).Setting up this mode requires acquiring personal API secret keys from openai.com or google.com. To enable you to experiment with different external service environments, Imagix 4D includes a complimentary tier of pre-allocated queries with its product license (50 for Gemini, 10 for OpenAI). To use your own organization's account, you must place your API strings into plain text files named AIKey_Google_Gemini or AIKey_OpenAI_GPT within the <imagix installation>/user/ai_keys directory. Administrators can also host these keys centrally in the <imagix installation>/data directory of an Imagix license server to automatically provision access for all active workstation clients across the network. Local AI Server (Self-Hosted Infrastructure)For organizations with stringent intellectual property or compliance policies, the Local AI Server mode ensures that 100% of your source code data remains isolated within hardware you control. Under the hood, this framework uses an open-source inference engine optimized to run quantized Large Language Models locally. The user interface mirrors the cloud experience, but includes a dedicated Local menu on the AI Analysis window’s toolbar, allowing developers to manually cancel running queries, reset context thresholds, or shut down the server process. Because this approach relies entirely on local compute resources rather than distributed cloud clusters, queries on standard desktop hardware can take several minutes to process compared to seconds for the other two modes.Setting up a local infrastructure requires downloading compatible model files (specifically optimized for deep code processing) and deploying them into the <imagix installation>/user/ai_models subdirectory. Links to recommended GGUF model files are available at www.imagix.com/download/ai_models.html. If a developer chooses to run the server locally on their own workstation (localhost), Imagix 4D manages the local server lifecycle automatically, starting it as a background subprocess that safely exits when the main GUI session ends. For faster performance or multi-user environments, administrators can configure a standalone network host machine. This requires executing a certificate generation utility to build secure TLS handshakes between the network clients and the host daemon (imgxaisrv). On Windows platforms, the local engine can be permanently registered as a Windows System Service to remain active across machine reboots. Interactive Chat-Based AI AnalysisThe interactive, chat-based approach serves as a zero-cost, manual bridge that allows you to leverage external browser-based chatbots or existing enterprise web subscriptions. Functionally, when you trigger an AI query, after Imagix 4D generates the package of target-specific information about your software, it is instantly copied to your local system’s clipboard. The user interface provides clear next-step guidelines directly in the main analysis pane, alerting you that the query is ready. You then navigate to your preferred web chatbot (such as ChatGPT or Claude) in an external browser, paste the query, and click “Send”.The AI generated answer then appears in the chatbot. The chat-based approach supports importing the AI answers back into the Imagix 4D environment, providing both integrated visibility and later reference. To streamline the import, Imagix 4D utilizes an automatic clipboard monitoring engine. The moment you click the chatbot’s native "Copy Response" button, Imagix 4D intercepts the text block from the system clipboard and immediately populates it into the Imagix 4D AI Analysis window. This mode requires absolutely zero technical setup or API key integration, making it ideal for immediate deployment or utilizing self-hosted corporate web chat tools. Comparison of AI Integration Modes
|