@tomer6
Here's a step-by-step guide on how to use the various features of Google AI Studio as shown in the video: Navigate the Main Interface: On the left sidebar, you'll find the main sections: Chat, Stream, Video Gen, and Starter Apps. There's also a History section to revisit previous interactions. Using the Chat Area: Model Selection: агрегированный ответ в виде временных меток Click the dropdown menu in the top right (under "Run settings") to choose an AI model (e.g., "Gemini 2.5 Pro Preview 03-25"). Hover over a model name in the dropdown. A pop-up will appear showing: Input/Output API Pricing. Best for: (e.g., Coding, Reasoning, Multimodal understanding). Use case: (e.g., Reason over complex problems, Tackle difficult code). Knowledge cutoff date (e.g., Jan 2025). Rate Limits: (e.g., 150 RPM, Free 5 RPM 25 req/day). Latency indicator. Understanding Model Types: Gemini series (2.5 Pro, 2.5 Flash, 2.0, 1.5): General-purpose models, some with "thinking modes" for longer reasoning. Gemma: Smaller models, often for developers and on-device capabilities. Other (LearnLM): Models designed for teaching and breaking down complex topics. Run Settings: Token count: Displays the current token usage. Temperature: Adjusts the creativity of the response. Tools (Dropdown): Grounding with Google Search: Toggle this on to allow the AI to use Google Search for up-to-date information. It will also provide search sources. (As shown with the boxing match example at 2:42). Structured output: Enable to get responses in formats like checklists or tables. Code execution: Allows Gemini to use code to solve complex tasks. Function calling: (More advanced) Lets you define functions that Gemini can call. Advanced settings (Dropdown): Safety settings: Click "Edit" to adjust how strictly the model filters content related to Harassment, Hate, Sexually Explicit, and Dangerous Content. You can choose to block most, block some, block few, or turn off filters. Output length: Set the maximum number of tokens in the response. Interacting with the Chat: System Instructions: Click the shield icon at the top of the chat prompt area to set optional tone and style instructions for the model (e.g., "You are a blunt business advisor..." at 5:41). Rerun Response: If you don't like an output, click the Gemini (three stars) icon on the response to have the model try again. View Model Thoughts: For "thinking" models, you can expand the "Thoughts (experimental)" section to see the model's step-by-step reasoning process. Compare Mode: Click the "Compare mode" icon (two arrows pointing away from each other) at the top of the chat interface. This splits the screen, allowing you to select two different models and send them the same prompt to compare their outputs. (Demonstrated at 4:20 with an email prompt). Importing Files: Click the plus icon (+) next to the "Type something" input field. You can choose to import from: My Drive, Upload File, Record Audio, Camera, YouTube Video, or Sample Media. YouTube Video Analysis: Select "YouTube Video," paste the URL, and then ask questions about the video's visual content (e.g., "What colour was the porsche?" at 9:10). Note: This analyzes visual content, not transcripts. For transcripts, use the main Gemini app with the YouTube extension. Audio Analysis: Select "Upload File" or "Record Audio." Once uploaded (e.g., "Sam altman NTVIEr.mp3" at 10:41), you can ask the model to analyze or summarize the audio content. Sample Media: Use pre-loaded images or videos to quickly test the model's multimodal capabilities. Using the Stream Area (Real-time Co-pilot): Click on Stream in the left sidebar. This feature allows Gemini to act as a digital co-pilot by observing your screen in real-time. Click Share Screen. You can then talk to Gemini (using your microphone) and ask it questions about what's on your screen. For example, while viewing a research paper, you can ask, "What is this bit mean here?" (As shown at 11:47). Gemini will respond by voice and text. Using the Video Gen Area: Click on Video Gen in the left sidebar. Model: Currently shows "Veo 2". Settings: Number of results: Adjust the slider (currently limited to 1 or 2). Aspect ratio: Choose between 16:9 (desktop) or 9:16 (mobile). Video duration: Select from 5s to 8s. Frame rate: (e.g., 24 fps). Resolution: (e.g., 720p). Negative prompt: Define what you don't want to see in the video. Generating Video: Describe your desired video in the text prompt. Alternatively, click the image icon to upload an image or select from sample media (e.g., the sea turtle at 14:11). Add a text prompt to animate the image (e.g., "Make the turtle swim"). Click Run. Once generated, you can play the video, download it, or export it to Google Drive. Using Starter Apps: Click on Starter Apps in the left sidebar. This section provides pre-built demo applications that showcase Gemini's capabilities (e.g., Video Toys, Gemini 95, Maps Planner, Flashcard Maker). Click on an app (e.g., "Flashcard Maker" at 15:34) to use it. Enter a topic (e.g., "AI terms") and click "Generate Flashcards." You can also click + Create an app to build your own using various templates like Empty, Prompt, Live, Chat example, Embeddings, Streaming, Code execution, etc. Accessing History: Click on History in the left sidebar to see a list of all your previous chats and prompts. You can click on any item to reopen and continue that conversation.
@DiceDecides
what happened to the image editor? I can't find it anymore and the new models can't do it
@microsatellites
Thank you for the latest updates.
@ttdsi5745
Why do I have 500 request per day? Ive been using this for 3 days straight now. Who's footing the bill?
@attorneyarrasmith
Make the turtle swim andnd just like that, I became an animation director.
@АртёмКутас-з8о
Why did they remove the AI training? Then you could make your own AI. I really liked it, but they removed it😒
@Deestroyer82
8th 💪
@patrikszvaigzne2287
Second!
@AC-425
Yo i'm first lol
@anikettiwari9147
Third
@perhaar
Third 😂
@S2Turbine