Account Registration & 7-Day Verification Lock
To maintain platform security and prevent automated abuse, Machan AI utilizes Google Firebase Authentication with a strict verification protocol:
- Registration: You can register via Email/Password, Google OAuth, or GitHub OAuth. Social logins are automatically verified.
- 7-Day Grace Period: If you register via Email/Password, you are granted a 7-day grace period to verify your email address via the link sent to your inbox.
- Verification Lockout: If your email remains unverified after exactly 7 days, your account will be soft-locked. You will be redirected to a "Verification Required" screen and barred from accessing the UI until verification is completed.
Guest Mode Limitations & Telemetry
Machan AI offers a "Guest Mode" for trial purposes, bypassing the need for immediate registration. However, it operates under strict infrastructural constraints:
- No Data Persistence: Guest chats are held in volatile memory. Navigating away, refreshing, or closing the tab will permanently erase your conversation.
- Feature Restrictions: Guests cannot access the Ceyon CDN Library, Project folders, image generation, or premium models (e.g., Claude, DeepSeek).
- Security Telemetry: To prevent bot spam, guest sessions log basic network metadata (IP-based region tracking via ipapi.co and User-Agent) to our Firebase backend. Rate limits are aggressively enforced.
AI Models, Puter.js Integration & "Insufficient Funds"
Machan AI functions as a routing interface, connecting you to multiple top-tier LLMs:
- Google Gemini (Default): Our native integrations (Gemini 2.5 Flash, 2.5 Pro) are highly optimized for speed and daily tasks.
- Puter.js Network (Premium Models): To offer models like Anthropic Claude, Meta Llama 4, and DeepSeek without charging subscription fees, we utilize the Puter.js "User-Pays" network.
- Billing & Errors: When using Puter-hosted models, you are consuming your personal Puter.com API credits. If you exhaust your free Puter quota, you will receive an "Insufficient Funds" error. Machan AI does not control this billing; you must visit your Puter Dashboard to add funds or switch back to our default Gemini models.
Document Canvas & AI Inline Editing
When the AI generates long-form text (essays, articles, or documents), it renders them inside a specialized Canvas Block.
- Rich Text Editing: You can click directly into the document to type, delete, or reformat text. It features its own isolated Undo/Redo stack (Ctrl+Z) that does not interfere with the rest of the application.
- AI Rewriting: Highlight any text inside the Canvas to summon the floating AI Toolbar. Type a command (e.g., "Make this sound more professional" or "Translate to Spanish"), and the AI will rewrite that specific highlighted section in place.
- Export Options: Click the download icon in the Canvas header to export the document instantly as a formatted PDF or a Microsoft Word (.doc) file.
- Fullscreen Mode: Click the expand arrows to snap the document into a distraction-free fullscreen editor.
Interactive AI Quizzes
Machan AI features a fully native, interactive Quiz module for learning and testing your knowledge.
- Activation: Click the "Interactive Quiz" quick-action button, or type `/quiz [Topic]` to trigger the quiz engine.
- Interface: The AI will generate a set of multiple-choice questions rendered as interactive cards. You can click to select answers, and the system will provide immediate visual feedback (Correct/Incorrect).
- Gestures: On touch devices, you can swipe left or right to navigate between quiz cards smoothly.
- Scoring: At the end of the quiz, you will receive a final score and the option to "Review Results" or "Retake" the quiz. Your score is saved to your chat history.
Email & Messaging Modules
When you ask the AI to draft an email or a text message, it generates a specialized UI block tailored for communication.
- Email Module: Features an editable "Subject" line and body. Clicking the Send icon opens a dropdown allowing you to instantly export the draft into Gmail, Outlook, Yahoo Mail, or your device's default Mail app.
- SMS Module: Formatted for short-form messaging. The export dropdown allows you to push the text directly to WhatsApp, Telegram, SMS, or your native device sharing menu (Web Share API).
File Uploads, Ceyon CDN & 128MB Library Limits
Machan AI features a robust file parsing system and persistent media Library:
- Upload Limits: You may attach a maximum of 10 files per prompt.
- Ceyon CDN Storage: When you upload a file or generate an AI image, it is securely hosted on the Ceyon Content Delivery Network. Every registered user is strictly capped at 128 MB of total lifetime storage.
- Managing Quotas: If you hit the 128 MB limit, AI image generation and new file uploads will fail to save to your Library. You must navigate to the Library tab and delete older files to free up space.
Google Drive Integration & OAuth Security
You can seamlessly import documents into Machan AI directly from your Google Drive:
- OAuth Token Handling: When you connect Drive, Google issues a temporary OAuth Access Token. We store this token in your browser's Local Storage and your Firebase account node to allow the Google Picker UI to render.
- Data Access: We only request "readonly" scope. Machan AI only downloads the specific files you manually click and select. We do not scan or index your Drive.
- Revocation: Tokens naturally expire. You can also manually revoke Machan AI's access at any time via your Google Account Security settings.
Projects & Workspace Organization
To keep your workspace clean, you can group chats into dedicated folders called Projects.
- Creating a Project: Click "Projects" in the sidebar, then click "New". You can assign a custom icon and color to your project folder.
- Moving Chats: Click the three dots (...) on any chat in the sidebar and select "Move to Project" to assign it to a workspace.
- Project Hub: Clicking on a Project opens the Hub, isolating only the chats relevant to that topic. Creating a new chat while inside a Hub automatically assigns it to that project.
- Deletion Warning: Deleting a project will give you a checkbox option to either delete just the folder (sending chats back to the main list) or to permanently delete the folder and all chats contained within it.
Slash Commands, @ Mentions & Keyboard Shortcuts
Power users can navigate Machan AI rapidly using keyboard inputs:
- Slash Commands (/): Type
/in the chat box to open the command menu. You can instantly trigger actions like `/voice` (dictate), `/model` (change AI model), `/temp` (start temporary chat), or `/quiz`. - @ Mentions: Type
@in the chat box to search your Library. You can quickly attach previously uploaded files or generated images without leaving the keyboard. - Global Shortcuts: Press
Ctrl + Shift + O(or Cmd+Shift+O on Mac) to start a new chat. PressCtrl + Kto open the global chat search. PressCtrl + /to view the full list of available shortcuts.
Voice Dictation (Web Speech API)
You can dictate prompts using your device's microphone by clicking the microphone icon or typing `/voice`.
- Processing: Machan AI utilizes the native Web Speech API built into your browser. We do not record, store, or transmit raw audio files to our servers.
- Android Deduplication: We have built-in logic to detect and fix transcription duplication bugs commonly found in Android's native speech recognition engine, ensuring clean text output.
- Visualizer: A dynamic waveform visualizes your voice input in real-time. Click the stop button to commit the text to the prompt box.
Web Search & URL Fetching Proxies
Machan AI can bypass standard LLM knowledge cutoffs by interacting with the live internet:
- Google Custom Search: Automated web searches utilize Google APIs to fetch the top 8 recent articles, synthesizing them into a cited response.
- URL Fetching (Proxies): If you ask the AI to summarize a specific link, the request is routed through external proxy services (Jina AI, CorsProxy.io, or AllOrigins) to bypass browser CORS restrictions.
- Security Warning: Never ask the AI to fetch or summarize URLs containing private data, passwords, authenticated sessions, or local hostnames, as these requests traverse public proxy infrastructure.
Live Financial, Currency, & Weather Data Disclaimers
The AI is equipped to trigger special programmatic tools for real-time widget rendering:
- Data Sources: Financial charts are powered by Yahoo Finance APIs; weather data by Open-Meteo; and currency rates by ExchangeRate-API.
- Liability Disclaimer: All live data is provided strictly for informational purposes. It may be delayed, inaccurate, or incomplete. Under no circumstances should Machan AI be used to execute financial trades or make critical safety decisions. Machan AI assumes zero liability for damages resulting from data inaccuracies.
Code Generation & Client-Side Sandbox Execution
Machan AI is highly optimized for software development:
- Exporting: All code blocks support direct 1-click copying and downloading as native files (e.g., .py, .js, .cpp).
- Live HTML Sandbox: If the AI generates HTML/CSS/JS, you can click "Run Preview" to render the code inside a sandboxed `iframe`.
- Execution Warning: Because AI can occasionally generate infinite loops or memory-intensive scripts, executing AI-generated code in the sandbox is done entirely at your own risk. It may cause browser freezing or crashes if the generated logic is flawed.
Temporary Chat (RAM-Only Mode)
For highly sensitive queries, users can toggle "Temporary Chat" via the glasses icon in the top header.
- How it works: The conversation is held entirely within your browser's volatile memory (RAM). The prompts are sent to the AI API, but they are never written to our Firebase Realtime Database.
- Data Destruction: The moment you refresh the page, close the tab, or toggle the mode off, the entire context is instantly and permanently destroyed. It cannot be recovered.
Feedback, Telemetry, & Moderation
To continuously improve the routing and stability of the app, we utilize a robust feedback system:
- In-Line Reactions: Clicking the Thumbs Up/Down icons logs a rating for that specific AI response to our database.
- Diagnostic Reports: If you use the "Report a Bug" or "Feedback" forms, the system packages your typed message alongside your current device telemetry (User-Agent, resolution, language) and a text snapshot of the active chat. This is securely transmitted to our backend to assist developers in patching the fault.
If the technical documentation above does not resolve your issue, or if you need to submit a Data Subject Access Request (DSAR) / Account Deletion request, please use the secure routing form below. Our engineering team will review the diagnostic logs attached to your submission.