> ## Documentation Index > Fetch the complete documentation index at: https://docs.usevelo.ai/llms.txt > Use this file to discover all available pages before exploring further. # Write a Prompt > Transform a static PDF into a narrated, scene-based video by uploading your file and providing context for the AI agent. Velo allows you to transform static PDF documents into narrated, scene-based videos. By uploading a file and providing specific context, the AI agent generates a structured script and voiceover tailored to your target audience. ## Initial File Upload **Step 1:** From the main dashboard, select **Upload a recording or PDF**. **Step 2:** Choose the desired PDF file from your local machine. Upload selection screen The upload interface for MP4, MOV, and PDF files. ## Configuring AI Instructions Once the PDF loads, you must define the scope and tone of the explanation. **Step 3:** Choose which pages to include. You can select **All pages** or specify a range. **Step 4:** Enter **Instructions for the agent** to set the professional tone. For example, you can prompt the agent to "Explain this PDF like a customer success engineer." **Step 5:** Select **Continue with prompt** to allow the system to analyze the document content. Screenshot 2026 05 28 At 12 00 51 AM Defining the page range and agent instructions. ## Defining Persona and Audience To ensure the generated voiceover resonates with viewers, the system requires specific setup details. **Step 6:** Select a persona for the voice to carry, such as **Customer Support**, **Founder**, or **Product Manager**. **Step 7:** Identify the primary audience, such as **Customers**, **Internal team**, or **Management**. **Step 8:** Provide any additional context if necessary, then click **Submit**. Persona selection list Persona selection list Choosing a persona to drive the script's voice and tone. ## Voice Selection and Script Review After the initial setup, you can customize the audio and review the auto-generated scenes. **Step 9:** Choose between a **Public Voice** or a personalized **Voice Clone**. **Step 10:** Review the generated script. The AI divides the content into distinct scenes. **Step 11:** Edit the narration text for any scene that requires adjustment. Script editor interface Script editor interface The scene-by-scene script editor for fine-tuning narration. ## Finalizing Video Generation Once the script and scene breaks are approved, initiate the final production phase. **Step 12:** Click the "**Go, Make my Velo"** button at the bottom of the editor. **Step 13:** Monitor the progress bar as the system matches the new audio to the visual scenes. Generation button Generation button Final confirmation to begin video generation. The result is a fully produced video message ready for sharing with your audience.