All Cloud Tools

This is the complete reference for all cloud tools available in Karma One. Tools are grouped by category. Each tool is available to the AI during conversations and can be invoked through natural language prompts.

Communication and Scheduling

Tools for managing reminders, notifications, memory, web search, and email.

| Tool | Description | |------|-------------| | create_schedule | Create a scheduled task with a cron expression or delay timer. Use for reminders, recurring tasks, and automated workflows. | | memory_manage | Save, search, or delete persistent memory entries that carry across conversations. Useful for storing preferences, notes, and reference data. | | web_search | Search the web and return results with links, snippets, and metadata. Supports domain filtering for targeted searches. | | send_email | Send an email via your configured email account. Supports multiple recipients, CC, BCC, and file attachments. | | send_notification | Send a push notification to the user's device with a title, message, and optional sound. | | text_to_speech | Convert text to spoken audio. Supports multiple voices and languages for generating audio content. |

Example prompts:

"Remind me to check the report every Monday at 9 AM."
"Remember that my preferred programming language is TypeScript."
"Search the web for the latest news about renewable energy."
"Send an email to team@example.com with the meeting notes attached."

Content Generation

Tools for creating images, videos, UI designs, and multimedia content.

| Tool | Description | |------|-------------| | generate_image | Generate a high-quality image from a text description using Gemini 3 Pro Image model. Supports multiple aspect ratios (1:1, 16:9, 9:16, etc.) and resolutions (1K, 2K, 4K). | | generate_video | Generate a short video (5-20 seconds) from a text description. Choose from cinematic, realistic, animated, artistic, documentary, promotional, or abstract styles. | | film_teaser_start | Start generating a film-style teaser video combining a user photo with a movie theme. Async operation that takes 10-15 minutes. | | film_teaser_status | Check the progress and status of a film teaser generation task. Returns percentage complete and download link when finished. | | generate_ui | Generate a complete UI design from a text description. Returns production-ready HTML/Tailwind/React code. Supports web, iOS, and Android platforms. | | refine_design | Make modifications to an existing UI design. Preserves the original and creates a new variant with the requested changes. | | apply_theme | Transform the visual appearance of a design by applying a preset theme (light, dark, minimal, vibrant, corporate, startup, iOS, material). | | export_design | Export a generated UI design to various formats: HTML, Tailwind, React, Vue, React Native, Figma, PNG, or PDF. | | clone_ui | Recreate a UI from a screenshot. Three fidelity modes: exact (pixel-perfect), semantic (structure-focused), adaptive (component library optimized). |

Example prompts:

"Generate an image of a mountain landscape at sunset in 16:9 aspect ratio."
"Create a 10-second cinematic video of waves crashing on a rocky shore."
"Design a login page with email and password fields, dark theme."
"Clone this screenshot as a React component using shadcn/ui."

Knowledge Management

Tools for uploading, processing, and querying documents and knowledge bases.

| Tool | Description | |------|-------------| | knowledge_base_search | Search across one or more knowledge bases. Supports Gemini File Search, Bubble RAG, and BFS (Bubble File Search) providers. Returns relevant passages with citations. | | file_upload | Upload a document for processing and future retrieval. Supports PDF, DOC, DOCX, PPT, PPTX, XLS, XLSX, TXT, MD, CSV, JSON, images, audio, and video. | | file_query | Query a specific uploaded file with a natural language question. Returns relevant sections from the document. | | file_status | Check the processing status of an uploaded file. Files go through: uploaded, parsing, parsed, chunked, summarizing, summarized, indexing, ready. | | file_list | List all uploaded files in the current session with their doc IDs, file names, statuses, and chunk counts. |

Example prompts:

"Search my knowledge base for information about the Q4 revenue forecast."
"Upload this PDF and tell me when it's ready for questions."
"What does the uploaded contract say about termination clauses?"

Code Execution

Tools for running code and managing cloud sandbox environments.

| Tool | Description | |------|-------------| | code_interpreter | Execute Python code in an isolated sandbox environment and return results including text output, errors, and generated files. | | sandbox_execute | Execute any shell command in the cloud sandbox. Pre-installed tools include Node.js, Python, Chromium, and common development libraries. | | sandbox_write_file | Create or overwrite a file in the cloud sandbox filesystem at a specified path. | | sandbox_read_file | Read the contents of a file from the cloud sandbox filesystem. | | sandbox_status | Check whether the cloud sandbox is running, idle, or stopped. Returns current resource usage and uptime. | | sandbox_release | Stop and clean up the cloud sandbox to free resources. Use when finished with code execution tasks. |

Example prompts:

"Run this Python script that analyzes the CSV data."
"Install pandas and matplotlib in the sandbox, then create a bar chart."
"Write a Node.js script that fetches data from this API."
"Release the sandbox -- I'm done with the coding task."

Stock Analysis

A comprehensive suite of 14 tools for equity research, technical analysis, and portfolio tracking.

| Tool | Description | |------|-------------| | stock_get_quote | Get a real-time stock quote including current price, daily change, volume, market cap, P/E ratio, EPS, bid/ask spread, and 52-week range. | | stock_get_history | Retrieve historical price data with configurable intervals (1m, 5m, 15m, 1h, 1d, 1w, 1mo) and custom date ranges. | | stock_get_chart | Generate a visual stock price chart with optional technical indicator overlays. Returns an image. | | stock_technical_analysis | Run technical indicators including RSI, MACD, Bollinger Bands, moving averages (SMA/EMA), Stochastic Oscillator, ATR, and OBV. | | stock_get_financials | Retrieve financial statements: income statement, balance sheet, and cash flow statement. Quarterly and annual data available. | | stock_news_sentiment | Analyze recent news articles and social media for sentiment (bullish/neutral/bearish). Returns sentiment score and key headlines. | | stock_compare | Compare multiple stocks side-by-side on valuation, performance, profitability, growth, and dividend metrics. | | stock_screener | Filter the entire market using criteria such as market cap, sector, P/E, dividend yield, revenue growth, and profit margin. | | stock_analyst_ratings | Get Wall Street analyst consensus ratings (buy/hold/sell), price targets, and recent rating changes. | | stock_options | Retrieve options chain data including strike prices, expiration dates, implied volatility, open interest, and Greeks (delta, gamma, theta, vega). | | stock_holdings | Track and manage a personal watchlist or portfolio. Monitor performance, allocation, and gains/losses. | | stock_sector_performance | View performance breakdown by market sector (technology, healthcare, energy, financials, etc.) over various timeframes. | | stock_market_movers | Get the top gainers, top losers, and most actively traded stocks for the current trading session. | | stock_insider_trades | View recent insider buying and selling activity for a specific stock or across the market. |

Example prompts:

"Get me a quote for AAPL and MSFT."
"Run a full technical analysis on NVDA."
"Compare Tesla, Rivian, and Lucid on valuation and growth metrics."
"Screen for large-cap stocks with P/E under 20 and dividend yield over 3%."
"What are the top market movers today?"
"Show me insider trades at Apple over the past 3 months."

Browser Automation

Tools for controlling a Playwright-powered browser to navigate websites, interact with elements, and extract data.

| Tool | Description | |------|-------------| | browser_navigate | Navigate the browser to a specified URL. Supports HTTP and HTTPS. | | browser_navigate_back | Go back to the previous page in the browser's navigation history. | | browser_screenshot | Take a screenshot of the current page viewport, full page, or a specific element. Supports PNG and JPEG formats. | | browser_snapshot | Capture the accessibility tree of the current page. Returns a structured text representation of all visible content and interactive elements. | | browser_click | Click on a page element identified by its accessibility reference. Supports left/right/middle click, double-click, and modifier keys. | | browser_type | Type text into an editable element (input, textarea). Can type slowly to trigger key handlers or submit with Enter. | | browser_fill_form | Fill multiple form fields at once. Handles text inputs, checkboxes, radio buttons, combo boxes, and sliders. | | browser_evaluate | Execute a JavaScript expression or function on the current page and return the result. | | browser_press_key | Press a keyboard key or key combination (e.g., Enter, Escape, Ctrl+A, ArrowDown). | | browser_select_option | Select one or more options from a dropdown/select element by value. | | browser_hover | Move the mouse cursor over an element to trigger hover states and tooltips. | | browser_drag | Perform a drag-and-drop operation between two elements on the page. | | browser_file_upload | Upload a file through a file input element or file chooser dialog. | | browser_tabs | Manage browser tabs: list all open tabs, create new tabs, close tabs, or switch between them. | | browser_wait_for | Wait for specific text to appear on the page, text to disappear, or a specified duration to elapse. | | browser_network_requests | List all network requests made since the page loaded. Filter by resource type (XHR, fetch, script, etc.). | | browser_console_messages | Retrieve browser console messages (log, error, warning, debug) for debugging. | | browser_resize | Resize the browser window to specific width and height dimensions. | | browser_handle_dialog | Handle browser dialogs (alert, confirm, prompt). Accept or dismiss, with optional prompt text input. |

Example prompts:

"Go to news.ycombinator.com and list the top 10 stories."
"Take a full-page screenshot of our landing page."
"Fill in the contact form with name 'John' and email 'john@example.com'."
"Open three tabs and search for 'AI trends' on Google, Bing, and DuckDuckGo."

System Integration (macOS)

Tools for interacting with the local macOS operating system. Requires the Karma One desktop application.

Screenshots and Clipboard

| Tool | Description | |------|-------------| | mac_screenshot | Capture a screenshot of the desktop. Modes: fullscreen, specific window, or user selection. Optional delay timer. | | mac_clipboard_read | Read the current text contents of the macOS clipboard. | | mac_clipboard_write | Write text content to the macOS clipboard for pasting elsewhere. |

Notifications and Speech

| Tool | Description | |------|-------------| | mac_notification | Display a macOS system notification with title, message, optional subtitle, and sound. | | mac_say | Use macOS text-to-speech to speak text aloud. Configurable voice and speech rate. | | mac_open_app | Launch a macOS application by name (e.g., "Safari", "Finder", "Terminal"). | | mac_open_url | Open a URL in the default web browser. |

Calendar

| Tool | Description | |------|-------------| | mac_calendar_list | List all calendars configured in Calendar.app. | | mac_calendar_events | Retrieve events within a date range. Filter by specific calendar. Returns title, time, location, and calendar name. | | mac_calendar_create | Create a new calendar event with title, start/end time, location, notes, and optional alarm. | | mac_calendar_delete | Delete a calendar event by matching title and start date. |

Mail

| Tool | Description | |------|-------------| | mac_mail_send | Send an email via Mail.app. Supports multiple recipients (comma-separated), CC, BCC, and file attachments. Sends immediately. | | mac_mail_list | List recent messages from a mailbox. Returns subject, sender, date, and read status. Filter to unread only. | | mac_mail_read | Read the full content of a specific email message by index. | | mac_mail_search | Search emails by subject or sender keywords. | | mac_mail_act | Perform actions on an email: mark as read, mark as unread, delete, or move to another mailbox. |

Safari

| Tool | Description | |------|-------------| | mac_safari_open | Open a URL in Safari. Default opens in a new tab. | | mac_safari_tabs | List all open tabs across all Safari windows with URL and title. | | mac_safari_read | Read the text content of a Safari tab (extracts innerText, max 10K characters). | | mac_safari_js | Execute JavaScript code in a Safari tab and return the result. | | mac_safari_close | Close a specific Safari tab by window and tab number. |

Finder

| Tool | Description | |------|-------------| | mac_finder_list | List files and folders in a directory. Returns name, path, type, size, and modification date. | | mac_finder_info | Get detailed information about a file or folder (kind, size, creation date, modification date). | | mac_finder_move | Move a file or folder from one location to another. | | mac_finder_copy | Copy a file or folder to a new location. | | mac_finder_create_folder | Create a new folder at a specified path. |

System

| Tool | Description | |------|-------------| | mac_system_info | Get system information: hostname, macOS version, CPU, memory, disk usage. | | mac_system_run | Execute a shell command on macOS. Commands may require user approval based on security policy. | | mac_applescript | Execute AppleScript code for advanced macOS automation (controlling apps, system preferences, etc.). |

Example prompts:

"Take a screenshot of my desktop."
"What's on my calendar for tomorrow?"
"Create a meeting called 'Team Sync' for Thursday at 2 PM."
"Read my latest unread emails."
"Send an email to boss@company.com with the subject 'Weekly Report'."
"List all files on my Desktop."

AI and Analysis

Tools for leveraging additional AI models and exporting content.

| Tool | Description | |------|-------------| | gemini_query | Query Google Gemini AI for general questions, conversations, and text generation. Supports multiple Gemini model variants and conversation continuity via session IDs. | | gemini_analyze_code | Analyze code using Gemini AI. Four analysis modes: explain (what code does), review (find issues), optimize (suggest improvements), find-bugs (identify bugs). | | analyze_video_url | Analyze video content from a URL. Supports YouTube, Bilibili, and other platforms. Returns summary, key frames, transcript, or detailed analysis. | | document_export | Export conversation content or generated documents to PDF, Word (DOCX), or Excel (XLSX) format. |

Example prompts:

"Ask Gemini to explain quantum computing in simple terms."
"Analyze this Python function for potential bugs."
"Summarize this YouTube video: [URL]"
"Export this conversation to PDF."

Camera and Location

Tools for capturing photos, video, and accessing device location.

| Tool | Description | |------|-------------| | camera_snap | Capture a photo from the device camera (macOS). Includes configurable warmup time for camera initialization. | | camera_clip | Record a short video clip from the device camera with configurable duration. | | screen_record | Record the screen on macOS with configurable duration and display selection. | | location_get | Get the current geographic location of the device (latitude, longitude). Requires location permission. |

Example prompts:

"Take a photo with my camera."
"Record a 10-second video clip."
"What's my current location?"

Sovereign AI

Tools that use on-premises or sovereign infrastructure for maximum data privacy. Data never leaves your controlled environment.

| Tool | Description | |------|-------------| | local_generate_image | Generate images using a locally deployed AI model on Karma Box. Fully on-premises; data never leaves your device. 3-10 second generation time. | | local_analyze_image | Analyze images using a locally deployed vision model. Supports PNG and JPEG. Fully on-premises for maximum privacy. | | sovereign_generate_image | Generate images using cloud sovereign AI infrastructure you control. Data processed within your sovereign boundary. | | sovereign_analyze_image | Analyze images using cloud sovereign AI vision model. Data processed within your sovereign boundary. |

Sovereign AI tools are activated with specific phrases: "Use the Advanced Sovereign AI model" (local) or "Use the Sovereign AI model" (cloud).

UI Design (Advanced)

Extended tools for UI design workflows, previewing, adapting, and design system management.

| Tool | Description | |------|-------------| | vision_preview | Generate an interactive preview of a UI design. Supports device frames (iPhone 14, Pixel 7) and grid overlays. | | vision_auto_push | Push a design preview to the right-panel Preview Tab for immediate viewing. Supports hot update for live changes. | | vision_hot_update | Send incremental updates to an active preview session without full page reload. | | vision_responsive | Generate responsive CSS with media queries. Supports mobile-first and desktop-first strategies with configurable breakpoints. | | vision_adapt_library | Adapt a design to a specific component library: shadcn/ui, Ant Design, MUI, Element Plus, or Chakra UI. TypeScript output supported. | | vision_export_tokens | Export design tokens in CSS, SCSS, Tailwind config, JSON, Style Dictionary, or Figma Tokens format. | | vision_get_assets | Browse design assets: icons (searchable), gradients, patterns, fonts, and placeholder images. | | vision_history | View the full version history of generated designs with tree visualization showing branching and parent-child relationships. | | vision_extract_design | Analyze a screenshot and extract its design language: colors, typography, spacing, shadows, icons, and design philosophy. | | vision_fuse_design | Intelligently blend the design language of a reference screenshot into a target design. Configurable intensity (subtle, moderate, aggressive). | | vision_compare_designs | Compare two design screenshots and get similarity scores across dimensions (color, typography, layout, etc.). |

Process and Task Management

Tools for managing background processes and scheduled jobs.

| Tool | Description | |------|-------------| | process_start | Start a shell command as a background process. Returns a process ID for monitoring. | | process_list | List all active and completed background processes with their status. | | process_poll | Poll for new output from a running background process. Supports reading from a specific line number. | | process_log | Get the complete log output from a background process. | | process_write | Write input text to the stdin of a running background process. | | process_kill | Terminate a background process with SIGTERM or SIGKILL signal. | | cron_create | Create a recurring scheduled job with a cron expression and a command to execute. | | cron_list | List all configured cron jobs with their schedules, commands, and enabled/disabled status. | | cron_delete | Delete a cron job by its ID. | | cron_toggle | Enable or disable a cron job without deleting its configuration. |

Example prompts:

"Start a background process that monitors the log file."
"Create a cron job that runs a backup script every night at midnight."
"Show me all running background processes."
"Kill process abc-123."

Quick Reference: Tool Count by Category

| Category | Count | |----------|-------| | Communication and Scheduling | 6 | | Content Generation | 9 | | Knowledge Management | 5 | | Code Execution | 6 | | Stock Analysis | 14 | | Browser Automation | 19 | | System Integration (macOS) | 28 | | AI and Analysis | 4 | | Camera and Location | 4 | | Sovereign AI | 4 | | UI Design (Advanced) | 11 | | Process and Task Management | 10 |

Notes

Not all tools are available in every context. Tool availability depends on your subscription plan, the platform you are using (web, mobile, desktop), and your permission settings.

macOS system integration tools require the Karma One desktop application and appropriate system permissions (accessibility, calendar, mail access).

Sovereign AI tools require either a locally deployed Karma Box or access to sovereign cloud infrastructure. These tools are designed for organizations with strict data residency requirements.

Tools that modify state (send email, create events, execute commands) respect your permission mode settings. In "Ask Every Time" mode, the AI will always request confirmation before invoking these tools.

You can invoke any tool through natural language. You do not need to know the tool name -- just describe what you want to do and the AI will select the appropriate tool automatically.