Gunzino Support

OpenIntelligence is an Apple Intelligence-fueled agentic RAG system for iPhone, iPad, and Mac. It combines Apple's Foundation Models framework, on-device model execution, PDFKit, Vision OCR, speech transcription, local indexing, citations, and a reusable engine boundary for grounded document work.

Document libraries, retrieval, citations, and answer verification stay on your device. On iOS, iPadOS, and macOS 27 and newer, a long evidence-heavy request can use Apple Private Cloud Compute for the final answer, and only after the app shows what would be sent and you approve it. Older OS versions run everything on device, and every answer shows the route that produced it.

OpenManual helps identify a device from its label, barcode, photo, or typed details, then find and review public manuals and companion documents that may apply. The app keeps selected documents locally and shows the source evidence behind answers.

OpenManual does not bypass logins, paywalls, CAPTCHAs, or other access controls. Obtain the document through the source’s normal authorized path, then import it through Files or Safari’s Share Sheet if you have the right to keep it. The dedicated OpenManual support page has the full import and deletion guidance.

No. OpenManual can help locate eligible device documentation, but it does not diagnose, recommend treatment, or choose patient-specific settings. Verify the applicable instructions for use and consult a qualified professional when needed.

OpenResponses is a native SwiftUI developer client for the OpenAI Responses API. It exposes direct request controls, SSE streaming events, reasoning traces, tool runs, computer-use approvals, and MCP or connector configuration in one app.

OpenResponses centers on the models and tool paths available to your OpenAI account, including GPT-5, GPT-4.1, compatible reasoning models, and the specialized computer-use paths exposed by the app. It fetches the live model list from OpenAI instead of relying on a hard-coded catalog.

Model Context Protocol (MCP) lets OpenResponses connect to remote MCP servers and certain hosted connectors, depending on your model, tool compatibility, and configured authentication. The app keeps approval flows visible so tool calls are not hidden behind a proxy.

Computer use is the app's browser automation harness for supported tool and model paths. It lets the model navigate sites, click, type, scroll, and capture screenshots, but every step stays behind visible approval and audit controls.

OpenCone is a cloud-hybrid RAG client. It parses and OCRs documents locally, sends chunk text to OpenAI embeddings, stores vectors in Pinecone, and sends retrieved context to the OpenAI Responses API for grounded answers.

You need an OpenAI API key for embeddings and responses, plus a Pinecone API key and project details for vector storage and search.

OpenCone supports PDF, DOCX, TXT files, and images (with OCR capability using VisionKit).

Yes. OpenCone uses secure connections for API integrations, security-scoped bookmarks for file access, and follows best practices for data handling. Your documents and embeddings are stored in your own Pinecone account.

OpenAssistant was a native iOS dashboard for the OpenAI Assistants API, built around assistants, threads, runs, files, and vector stores.

OpenAssistant is an archived product line. Ongoing work moved to OpenResponses, which targets the newer direct Responses API model instead of the older Assistants API thread-and-run workflow.

Availability varies by product. Some lines ship or shipped through the App Store, OpenClinic remains a prototype, and OpenAssistant is archived. Use each product page for the current download and support status instead of assuming the whole catalog works the same way.

Contributions are welcome! Visit the GitHub organization repositories, fork the code, and submit pull requests. Check each project's README for contributing guidelines.

Please open an issue on the respective application's GitHub repository. For general feedback, use the contact form above.