# Media AI Chat *** ## Table of Contents * [Overview](#overview) * [Chat Methods](#chat-methods) * [Method 1: Chat with Projects](#method-1-chat-with-projects) * [Method 2: Chat with Projects & Team](#method-2-chat-with-projects--team) * [Method 3: Chat with Attachments](#method-3-chat-with-attachments) * [Method 4: Chat in Media Manager](#method-4-chat-in-media-manager) * [Supported Media Types](#supported-media-types) * [Use Cases](#use-cases) * [Media & Workspace DNA](#media--workspace-dna) * [What's Next](#whats-next) *** ## Overview Media AI Chat lets you **have conversations with your files, projects, and media** using Taskade's AI. Upload documents, images, videos, or connect projects — then ask questions, extract insights, and generate content directly from your media. > **Genesis connection:** Every file you upload to a Genesis app or discuss in Media AI Chat is stored in your workspace's Media library, enriching your [Workspace DNA](https://docs.taskade.com/docs/taskade-genesis/genesis/how-genesis-works) and making all future apps and agents smarter. *** ## Chat Methods ### Method 1: Chat with Projects Query your Taskade projects using natural language: | Step | Action | | ---- | ----------------------------------------------- | | 1 | Open a project in Taskade | | 2 | Click the **AI Chat** or **EVE** panel | | 3 | Ask questions about the project content | | 4 | AI responds using the project's data as context | **Example queries:** | Query | What AI Returns | | --------------------------------- | ------------------------------------------------- | | "Summarize this project" | A concise summary of all tasks, notes, and status | | "What tasks are overdue?" | List of past-due items with details | | "Who is assigned the most tasks?" | Team workload breakdown | | "Generate a status report" | Formatted report based on project data | ### Method 2: Chat with Projects & Team Collaborate with AI and team members in the same conversation: | Step | Action | | ---- | ------------------------------------------------------------ | | 1 | Open a project's chat panel | | 2 | Invite team members to the conversation | | 3 | Ask AI questions — responses are visible to all participants | | 4 | Team members can follow up, add context, or redirect the AI | > **Tip:** Use this for collaborative analysis sessions — the AI provides data, your team discusses and decides. ### Method 3: Chat with Attachments Upload files directly into a chat for instant AI analysis: | Step | Action | | ---- | ------------------------------------------------ | | 1 | Open any AI chat (EVE, Agent, or Project chat) | | 2 | Drag & drop files or click the attachment button | | 3 | Ask questions about the uploaded file | | 4 | AI analyzes the file content and responds | **Example interactions:** | File Type | Query | AI Response | | --------------- | ----------------------------------------- | -------------------------------------- | | **PDF report** | "What are the key findings?" | Summary of main points and conclusions | | **CSV data** | "What trends do you see?" | Pattern analysis and insights | | **Image** | "Describe what you see" | Visual analysis and description | | **Spreadsheet** | "Which products have the highest margin?" | Data analysis with specific answers | ### Method 4: Chat in Media Manager Access all workspace media in one place: | Step | Action | | ---- | ------------------------------------------------------------- | | 1 | Open the **Media** tab in your workspace | | 2 | Browse all uploaded files, generated images, and media assets | | 3 | Click any media item to open it | | 4 | Use AI chat to ask questions about the selected media | *** ## Supported Media Types | Media Type | Formats | AI Capabilities | | ----------------- | ----------------------------------------- | --------------------------------------------------- | | **Documents** | PDF, DOCX, TXT, MD, PPTX, XLSX, CSV, EPUB | Full text analysis, summarization, Q\&A, extraction | | **Images** | PNG, JPG, GIF, SVG, WebP | Visual description, content extraction, OCR | | **Video** | YouTube URLs (via transcript) | Transcript analysis, key points extraction, Q\&A | | **Web Content** | URLs, blog posts, articles | Content analysis, summarization, comparison | | **Spreadsheets** | XLSX, CSV, Google Sheets | Data analysis, trend identification, reporting | | **Presentations** | PPTX | Slide content analysis, outline extraction | | **Social** | Tweets, Reddit posts, news articles | Sentiment analysis, summary, response drafting | *** ## Use Cases | Use Case | How | Example | | ------------------------ | ----------------------------------------- | ------------------------------------------------------------------ | | **Document analysis** | Upload a report, ask questions | "What are the Q3 revenue figures from this financial report?" | | **Data exploration** | Upload a CSV, ask for patterns | "Which customer segment has the highest churn rate in this data?" | | **Content extraction** | Upload a PDF, extract structured data | "Extract all contact information from this business card scan" | | **Image understanding** | Upload a screenshot, get description | "What UI elements are shown in this mockup?" | | **Meeting prep** | Chat with project notes before a meeting | "Summarize the open items and blockers from last sprint's project" | | **Competitive analysis** | Add competitor URLs, compare | "Compare the features listed on these two competitor pages" | | **Knowledge building** | Upload training materials, quiz yourself | "Create 5 quiz questions from this training document" | | **Report generation** | Chat with data, generate formatted output | "Create a monthly report from this spreadsheet data" | *** ## Media & Workspace DNA Every media interaction enriches your workspace: | Media Action | DNA Impact | | --------------------------- | ---------------------------------------------------------------------- | | Upload files to Genesis app | Added to Media library → available for future apps and agent knowledge | | Generate images in chat | Saved to Media → reusable across all apps | | Upload agent knowledge | Indexed for agent responses → better AI across workspace | | Analyze documents | Insights inform future app builds | | Chat with project data | Helps EVE understand your workspace patterns | > **Cross-link:** Files uploaded in Media AI Chat can be used as [Genesis Context](https://docs.taskade.com/docs/taskade-genesis/genesis/adding-context), [Agent Knowledge](https://docs.taskade.com/docs/ai-features/ai-features/agent-knowledge), and as data sources for [Automations](https://docs.taskade.com/docs/automations/automation). *** ## What's Next | Guide | What You'll Learn | | ------------------------------------------------------------------------------------------------- | ------------------------------------------------- | | [Adding Genesis Context](https://docs.taskade.com/docs/taskade-genesis/genesis/adding-context) | Use media files as context for app building | | [Agent Knowledge & Memory](https://docs.taskade.com/docs/ai-features/ai-features/agent-knowledge) | Train agents on your media files | | [Image Generation](https://docs.taskade.com/docs/ai-features/ai-features/image-generation) | Generate new images to add to your media library | | [Taskade AI Credits](https://help.taskade.com/en/articles/taskade-ai-credits) | Understand credit usage for media AI interactions |