OpenResponses is a native SwiftUI app for iOS and macOS, designed as a powerful and flexible playground for exploring the full range of OpenAI's API capabilities. It provides a robust chat interface, deep tool integration, and extensive customization options for developers and power users aiming to work with the latest models and features.

This project is currently in a "super beta" state, with a focus on achieving 100% compliance with the latest OpenAI APIs, including advanced tools like computer and backend conversation management.

• Multi-Model Support: Switch between a wide range of models, categorized for clarity:
  • Latest Models: gpt-5, gpt-4.1, and their variants.
  • Standard Models: gpt-4o, gpt-4, and gpt-3.5-turbo.
  • Reasoning Models: o-series models for complex, multi-step tasks.
  • Specialized Models: computer-use-preview for browser automation tasks.
• Dynamic Model Selection: Automatically fetches and displays the latest available models from OpenAI's API with intelligent categorization.
• Enhanced Streaming Experience: Real-time responses with granular status updates showing when the AI is thinking, searching the web, generating code, or running specific tools.
• Cancellable Operations: Stop streaming responses mid-generation with full control.
• Native SwiftUI Interface: A clean, responsive, and platform-native experience for iOS and macOS.

• Computer Use: A production-ready tool (primarily for the computer-use-preview model) that allows the model to control a web browser to perform complex tasks. It supports a full range of actions including navigation, clicking, typing, and scrolling with robust error handling.
• Web Search: Access up-to-date information from the internet, with support for both standard and preview search capabilities.
• Code Interpreter: Execute Python code in a secure, sandboxed environment. Features include:
  • Container Selection: Choose between auto, secure, or gpu execution environments.
  • File Preloading: Provide file IDs to be available in the sandbox from the start.
  • Artifact Parsing: Richly displays outputs like logs, images, and data files.
• Image Generation: Create images from text prompts using gpt-image-1 with real-time streaming previews and configurable quality settings.
• File Search: Perform searches across multiple vector stores simultaneously, enabling powerful knowledge retrieval from uploaded documents.
• MCP Integration: Connect to Model Context Protocol (MCP) servers for extended functionality, featuring a discovery service for popular platforms and a secure approval workflow.
• Custom Functions: Define and call your own custom functions within the chat flow.

• Direct File Uploads: Upload files directly from your device for use in prompts, supporting over 43 file types.
• Vector Store Management: Organize and manage documents in vector stores for efficient file searching.
• Smart Organization: Group related documents for better search results.

• Prompt Library: Save, manage, and reuse custom prompt configurations.
• Preset System: Create named presets with all settings (models, tools, parameters).
• Quick Switching: Instantly switch between different conversation configurations.

• Core Parameters: Adjust temperature, max_output_tokens, presence_penalty, frequency_penalty, top_p, and more.
• Reasoning Configuration: Configure reasoning effort and summary output for O-series and reasoning models like gpt-5.
• Include Controls: Choose what additional data to include in API responses.
• Service Tier Management: Configure API service tiers and background processing.
• Parallel Tool Calls: Control concurrent tool execution for efficiency.

• API Inspector: View detailed API requests and responses for transparency and debugging.
• Debug Console: Real-time debug logs with filtering by level and category.
• Analytics Dashboard: Comprehensive request tracking and performance monitoring.
• Network Logging: Detailed request/response logging with JSON pretty-printing.

• Enhanced Streaming Feedback: Real-time typing indicators with animated cursors during message generation.
• Live Token Estimation: Per-message and conversation-level token counters that update during streaming.
• Activity Feed: Expandable activity panel showing real-time updates during streaming:
  • Tool execution status (web search, code execution, image generation)
  • Reasoning and model thinking phases
  • MCP server interactions and approvals
  • Computer use actions and safety checks
• Streaming Status Indicators: Granular status updates for different phases of response generation.

• MCP Tool Approval Workflow: User-in-the-loop approval system for MCP server tool calls with detailed request information.
• Safety Approval for Computer Use: Review and approve computer actions that may require user confirmation.
• Secure Credential Management: All API keys and sensitive data stored in iOS/macOS Keychain.

• Intelligent URL Detection: Automatic detection and extraction of URLs from messages using URLDetector.swift.
• Embedded Web Content: Rich web content rendering with WebContentView.swift using native WKWebView integration.
• Content Filtering: Smart filtering for renderable content with support for markdown images, data URLs, and sandbox paths.
• Ad Blocking: Built-in ad blocking and security restrictions for web content.
• Desktop User Agents: Automatic desktop user agent configuration for optimal web content display.

• Enhanced Image Display: Advanced image rendering with EnhancedImageView.swift featuring animations, fullscreen support, and save functionality.
• Image Processing Utilities: Optimized image handling with ImageProcessingUtils.swift for memory management and display optimization.
• Image Reconstruction: Robust image reconstruction from various formats with fallback handling.
• Memory Management: Smart image optimization to prevent memory issues with large images.
• Placeholder System: Beautiful placeholder images for loading states.
• Image Generation Suggestions: Smart suggestion system with ImageSuggestionView.swift that appears when users type image-related keywords.
• Quick Image Generation: One-tap image generation from the input toolbar.

• Multi-Page Onboarding: Comprehensive onboarding flow with OnboardingView.swift for first-time users.
• API Key Setup Guidance: Step-by-step API key configuration with visual guidance.
• Welcome Experience: Interactive introduction to tools and features.
• Progressive Disclosure: Gradual introduction of advanced features.

• Advanced Text Formatting: FormattedTextView.swift with full Markdown support including code blocks, bold, italics, and syntax highlighting.
• Code Interpreter Artifacts: Comprehensive artifact system with ArtifactView.swift supporting 43+ file types including logs, CSV, JSON, Python files, and archives.
• Expandable Content: Collapsible artifact views with copy functionality and proper MIME type handling.
• Message Bubbles: Rich message rendering with MessageBubbleView.swift supporting streaming indicators, formatted text, and media content.

• Visual File Previews: SelectedFilesView.swift provides thumbnail previews for attached files with proper file type icons.
• Multi-File Selection: Support for selecting and managing multiple file attachments simultaneously.
• File Type Recognition: Intelligent file type detection with appropriate icons for documents, images, archives, and code files.
• Image Gallery: ImagePickerView.swift with grid-based image selection and preview capabilities.

• Local Conversation Storage: Comprehensive conversation persistence using ConversationStorageService with JSON-based file storage.
• Conversation Export: Export conversations as formatted text for sharing.
• Conversation Switching: Seamlessly switch between multiple conversation threads.
• Message Management: Delete individual messages, clear conversations, and manage chat history.

• MCP Server Discovery: MCPToolDiscoveryView.swift provides an intuitive "app store" experience for browsing and configuring MCP servers.
• Built-in Server Registry: Curated collection of popular MCP servers including GitHub, Notion, Slack, Google Drive, Shopify, Airtable, Weather, and Calculator.
• Intelligent Search & Filtering: Semantic search across server names, descriptions, and tool capabilities with category-based filtering.
• Secure Authentication Management: Ultra-secure keychain storage for MCP server authentication with automatic migration from insecure storage.
• Interactive Approval System: MCPApprovalView.swift presents beautiful approval sheets with parsed tool arguments and security warnings.
• Real-Time Tool Monitoring: Complete integration status tracking with MCPIntegrationStatus.swift for comprehensive debugging.
• Per-Tool Configuration: Granular control over which tools are enabled for each MCP server with visual configuration interface.

• Full Accessibility Support: VoiceOver compatibility with proper labels, hints, and navigation.
• Dynamic Type Support: Adapts to user's preferred text sizes.
• Keyboard Navigation: Complete keyboard accessibility for all features.
• Error Recovery: Graceful error handling with user-friendly messages.
• Comprehensive UI Testing: Full UI testing suite with accessibility identifiers and launch screen testing.
• Internationalization Ready: Complete localization support with Localizable.xcstrings supporting multiple languages including Spanish.
• App Store Ready: Production-ready implementation with comprehensive app store assets, screenshots, and metadata preparation.
• Multi-Platform Support: Native support for iOS, iPadOS, macOS, and visionOS with platform-specific optimizations.

• API Request Inspector: Real-time API request/response monitoring with APIInspectorView.swift for complete transparency.
• Comprehensive Logging: Structured logging system with AppLogger.swift featuring categorized logs, OpenAI-specific logging, and streaming event tracking.
• Performance Analytics: AnalyticsService.swift tracks API performance, request sizes, response times, and error rates with detailed metrics.
• Network Monitoring: Real-time network connectivity monitoring with NetworkMonitor.swift and offline handling with graceful degradation.
• Debug Console: Advanced debugging interface with DebugConsoleView.swift showing system information and detailed logs.
• Token Usage Tracking: ConversationTokenCounterView.swift provides detailed token consumption and cost estimation with real-time updates.
• Request History: Complete API request history with request/response inspection for troubleshooting and debugging.
• Error Handling & Retry Logic: Sophisticated error handling with automatic retry mechanisms for transient failures and rate limiting.
• Rate Limiting Support: Built-in rate limiting detection with automatic backoff and retry-after header support.
• Alternative Service Implementations: Multiple OpenAI service implementations including minimal testing versions for development scenarios.

OpenResponses is built with SwiftUI and follows the MVVM (Model-View-ViewModel) pattern with feature-oriented folders that keep views, models, and services focused and discoverable.

• OpenResponses/App: Application bootstrap code such as OpenResponsesApp.swift, AppContainer.swift, and logging/analytics setup.
• OpenResponses/Core: Shared models, services, and protocols (ChatMessage.swift, OpenAIService.swift, APICapabilities.swift, KeychainService.swift, etc.).
• OpenResponses/Features: Feature-specific views and view models. For example, the chat feature lives in Features/Chat with UI components, ChatViewModel.swift, and related extensions like ChatViewModel+Streaming.swift that now own streaming handlers.
• OpenResponses/Shared: Cross-cutting UI components and utilities reused across multiple features.
• OpenResponses/Resources: Assets, localized strings, and UI resources (Assets.xcassets, Launch Screen.storyboard, Localizable.xcstrings).
• OpenResponses/Support: Legacy implementations, sample data, and developer tooling kept separate from production code.

• ChatView.swift: Primary conversation UI that renders message history, the input surface, and feature sheets.
• ChatViewModel.swift + ChatViewModel+Streaming.swift: The view model manages chat state, orchestrates API calls, and delegates streaming event parsing to the dedicated extension to keep functions small and focused.
• ContentView.swift: Coordinates onboarding, API key setup, settings presentation, and deep links.
• OpenAIService.swift: Central networking layer that builds /v1/responses payloads using composable helpers for models, tools, and attachments, and streams events back to the view model.
• ConversationStorageService.swift: Handles persistence of local conversations and metadata.
• AppContainer.swift: Dependency registry that wires services together for injection throughout the app.
• APICapabilities.swift & ModelCompatibilityService.swift: Type-safe definitions describing available tools and which models support them.

• Prompt & File Management: PromptLibrary and FileManagerView provide reusable prompt presets and file/vector store workflows.
• Diagnostics & Debugging: Views such as APIInspectorView, DebugConsoleView, and ConversationTokenCounterView aid in troubleshooting and monitoring usage.
• Tool Integrations: ComputerService, MCP discovery components, and code interpreter helpers live in Core/Services and are surfaced through feature modules.
• MCPIntegrationStatus.swift: Complete MCP integration status monitoring and validation system.

• ChatView.swift: The main chat interface with message display, input handling, and tool indicators.
• MessageBubbleView.swift: Individual message rendering with streaming support and rich content display.
• StreamingStatusView.swift: Real-time status indicators for different streaming phases with DotLoadingView animations.
• ActivityFeedView.swift: Expandable activity feed showing detailed streaming progress.
• ConversationListView.swift: Interface for browsing and managing conversation history.
• MCPApprovalView.swift: User approval interface for MCP tool calls.
• SafetyApprovalSheet.swift: Safety confirmation interface for computer use actions.
• WebContentView.swift: Rich web content rendering with embedded WKWebView support.
• URLDetector.swift: Intelligent URL detection and extraction from message content.
• EnhancedImageView.swift: Advanced image display with animations, fullscreen, and save capabilities.
• FormattedTextView.swift: Markdown text rendering with code block syntax highlighting.
• ArtifactView.swift: Code interpreter artifact display supporting 43+ file types.
• ImageSuggestionView.swift: Smart image generation suggestions based on user input.
• OnboardingView.swift: Multi-page onboarding flow for first-time users.
• ImagePickerView.swift: Grid-based image selection with preview capabilities.
• SelectedFilesView.swift: Visual file attachment previews with type recognition.
• ModelCompatibilityView.swift: Model capability and tool compatibility display.
• DynamicModelSelector.swift: Intelligent model selection with compatibility checking and real-time fetching.
• ConversationTokenCounterView.swift: Real-time token usage monitoring and display.

• DotLoadingView: Animated ellipsis indicator with fluid motion for streaming status.
• TypingCursor: Blinking cursor animation for real-time typing indication during streaming.
• CompactToolIndicator: Minimal tool status display with animated indicators for active tools.
• MessageToolIndicator: Individual message tool usage indicators with visual feedback.

• APIInspectorView.swift: Real-time API request/response inspection for debugging.
• DebugConsoleView.swift: Live debug log viewer with filtering capabilities.

Comprehensive Codable data structures providing full type safety and matching the OpenAI API's JSON structure:

• StreamingEvent.swift: Complete streaming event system supporting 40+ event types from the OpenAI Responses API.
• ChatMessage.swift: Core message model with support for text, images, tools, and token usage tracking.
• ResponseModels.swift: Complete response object models for all API interactions.
• ComputerModels.swift: Computer use tool models and action definitions.
• MCPModels.swift: Model Context Protocol integration models and server configurations.
• OpenAIModel.swift: Model listing and capability definitions.
• Conversation.swift: Local conversation storage model with metadata and persistence support.
• Prompt.swift: Configuration model containing all prompt settings, tool selections, and parameters.
• StreamingStatus.swift: Enumerated streaming states with user-friendly descriptions.

• macOS with Xcode installed.
• An OpenAI API Key from the OpenAI Platform.

1. Clone the repository:

git clone https://github.com/Gunnarguy/OpenResponses.git
cd OpenResponses

1. Open the project:

Open the OpenResponses.xcodeproj file in Xcode.

1. Build and Run:

Select your target device (iOS or macOS) and run the app.

1. Configure the App:

   • On first launch, navigate to the Settings screen.
   • Enter your OpenAI API key to enable communication with the API.
   • Select your preferred model and configure any desired tools or advanced settings.
   • Optional: Create prompt presets to save your favorite configurations for quick switching.

For detailed guides on all features, please refer to our comprehensive documentation:

• /docs/PromptingGuide.md: Best practices for writing effective prompts.
• /docs/Tools.md: A guide to using tools like Web Search, File Search, and Function Calling.
• /docs/Images.md: A guide to image generation and vision capabilities.
• /docs/Advanced.md: A guide to advanced features like streaming, structured outputs, and prompt caching.
• /docs/FILE_MANAGEMENT.md: A user guide for file and vector store features.
• /docs/NOTION_MCP_GUIDE.md: Complete guide to connecting Notion via Model Context Protocol.
• /API/StreamingEventsAPI.md: Complete documentation of all supported streaming events.

The app automatically fetches and categorizes the latest models from OpenAI, including:

• Latest Models: gpt-5, gpt-4.1 series with full tool support
• Reasoning Models: o-series models optimized for complex thinking tasks
• Specialized Models: computer-use-preview for browser automation This includes intelligent filtering for chat-compatible models and provides real-time descriptions to help you choose the best model for your task.

The app provides comprehensive streaming feedback through multiple layers:

• Streaming Event Processing: Handles 40+ different streaming event types from the OpenAI API
• Real-time Status Updates: Granular status indicators showing exactly what phase the model is in:
  • thinking for reasoning phases
  • searchingWeb for web search operations
  • generatingCode for code interpreter execution
  • generatingImage for image creation with progress updates
  • usingComputer for browser automation actions
  • mcpApprovalRequested for MCP tool confirmation
• Activity Feed: Detailed, user-friendly activity log showing:
  • Tool execution progress
  • Reasoning steps and model thinking
  • File processing and artifact generation
  • Network status and rate limiting information

The app features a complete MCP (Model Context Protocol) ecosystem:

• Discovery Service: Built-in registry of popular MCP servers (GitHub, Notion, Slack, etc.)
• One-click Setup: Streamlined configuration for known services
• Security-First: All credentials stored in Keychain with approval workflows
• Real-time Tool Discovery: Dynamic detection of available tools from MCP servers

Save and manage reusable prompt configurations in the Prompt Library. Configure your settings once, save it as a preset, and quickly switch between different setups for different tasks.

Upload, organize, and search through your documents using Vector Stores with multi-store search capabilities. For more information, see the File Management Guide.

• API Inspector: View detailed request/response data with full JSON formatting
• Debug Console: Real-time logs with filtering by category and severity
• Analytics Dashboard: Track API usage, performance metrics, and token consumption
• Streaming Event Viewer: Monitor all streaming events in real-time for development

This project includes comprehensive testing capabilities:

1. In Xcode, open the Test Navigator (Cmd+6).
2. Click the play button next to the OpenResponsesUITests test suite to run all tests.

• API Inspector: Test API calls and view real-time request/response data.
• Debug Console: Monitor app behavior with detailed logging.
• Accessibility Testing: Full VoiceOver support for testing accessibility features.

• See PRODUCTION_CHECKLIST.md for a comprehensive pre-release validation guide.
• Includes performance testing, error handling validation, and accessibility verification.

Contributions are welcome! If you'd like to contribute, please follow these steps:

1. Fork the repository.
2. Create a new branch (git checkout -b feature/your-feature-name).
3. Make your changes and commit them (git commit -m 'Add some feature').
4. Push to the branch (git push origin feature/your-feature-name).
5. Open a Pull Request.

This project is licensed under the MIT License. See the LICENSE file for details.

• /docs/ROADMAP.md: The strategic master plan for all features and implementation priorities.
• /docs/Tools.md: A guide to using tools like Web Search, File Search, and Function Calling.
• /docs/Images.md: A guide to image generation and vision capabilities.
• /docs/Advanced.md: A guide to advanced features like streaming, structured outputs, and prompt caching.
• /docs/PromptingGuide.md: Best practices for writing effective prompts.
• /docs/CASE_STUDY.md: A technical deep-dive into the app's architecture and design decisions.
• /docs/NOTION_MCP_GUIDE.md: Complete guide to connecting Notion via Model Context Protocol.
• /docs/api/Full_API_Reference.md: Field-level API implementation status.
• /docs/PRODUCTION_CHECKLIST.md: A comprehensive pre-release validation checklist.
• /docs/FILE_MANAGEMENT.md: A user guide for file and vector store features.
• /docs/APP_STORE_GUIDE.md: App Store submission guidance.
• /docs/PRIVACY_POLICY.md: The app's privacy policy.