Good Morning! Here's what I have for you in today's newsletter:
xAI lets Grok Bot use your phone’s Calendar, Contacts, and Reminders
Canva launches Craft, its first image model built for design
Anthropic brings the full Claude Platform to Google Cloud
Make a 3D product launch film with Opus 5.5 and sound
4 new AI tools worth trying today
AI AGENTS
xAI rolled out phone app access for Grok Bot on iOS and Android. Give it permission, and the Bot can work with built-in apps such as Calendar, Reminders, and Contacts.
Permissions are separate for each app. Calendar access lets it check your schedule and add events; Contacts lets it find and save people. The permission screen also lists Messages for texts and Home for lights and locks.
A 1Password integration rolling out over the next few days lets the Bot sign in to websites. You choose and approve the logins it can use, while passwords stay in 1Password.
Depending on your operating system, share selected contacts or your whole address book. You can revoke any app’s access in your phone’s settings.
“Add dinner on Friday” or “find Sam’s number” can become a request to your Bot instead of a trip through several apps. Website logins give it more room to help, too. Access is yours to control, so choose the apps and accounts it needs before putting it to work.
AI MODELS
Canva released Craft, its first image generation model, built by the company’s own research team. All Canva users can try it through the Dream Lab tab.
Craft works out the layout, proportions, visual hierarchy, and text before rendering. It supports up to 2K output across 30 aspect ratios, and Canva says the planning helps keep words readable.
Training included human-designed Canva posters, menus, and infographics. Change one part without altering the rest, or use Canva Magic to turn the image into an editable, layered design.
Craft entered DesignArena in fourth place, which Canva says puts its lab second for image generation. API access, ChatGPT and Gemini support through Canva’s MCP, and cheaper Fast and Ultrafast versions are coming.
A good-looking menu isn’t much use if the dish names are unreadable. Craft plans where the words and visuals belong before drawing, then lets you keep editing the result as a layered design. For anyone already making posts, posters, or menus in Canva, it’s right there to try.
AI TOOLS
Anthropic released Claude Models on Google Cloud (Anthropic managed) for all Google Cloud customers. It offers the full Claude Platform using Google Cloud IAM credentials and billing, with spending counting toward cloud commitments.
New features and betas arrive on release day, including Claude Managed Agents, skills, and code execution. Models include Claude Fable 5.1, Opus 5.5, Sonnet 5.5, and Haiku 5.5.
Anthropic operates the service and processes data under its Commercial Terms. Vertex AI remains a separate, Google-managed option, where Google processes the data and adds features after integrating them.
Fast Mode for Opus 5.5 is also generally available. Subscribe through Google Cloud Marketplace, with US and Global endpoints available.
Teams on Google Cloud can get new Claude features without waiting for Google to integrate them into Vertex AI, while keeping their existing cloud credentials and billing. The choice also changes who processes the data: Anthropic or Google. Check that distinction against your company’s agreements before switching.
HOW TO AI
Turn your brand into a 14-second launch film with 3D motion and sound.
Opus 5.5 builds the scene in code, renders it inside Claude, and adds synthesized sound, with colors, fonts, and components from your design system. You'll need an eligible paid plan, code execution, and your brand files.
Step 1: Open a render chat
Enable code execution under Settings, then Capabilities. Start a chat with Opus 5.5 and ask for a one-second test: a spinning Three.js box rendered with Playwright and encoded with ffmpeg. Open the MP4 before continuing. Nothing needs installing on your computer.
Step 2: Load your brand
Ask Claude to create brand.json from your design system or attached brand sheet. Include color roles, fonts, corner radii, component styles, and your logo. Have it check available fonts and list missing details. Approve substitutes before they appear throughout the film.
Step 3: Write the story
Fill in the full guide's six fields: PRODUCT, CHAOS, JOURNEY, STOPS, ACTION, and END. The example turns scattered trip plans into a route across a 3D map, then pulls back to reveal a phone. Ask for twelve stills before rendering. Check text, colors, and framing first.
Step 4: Render with sound
Approve the stills, then ask for all 420 frames and sound synthesized in Python. Keep picture and sound events in timeline.py so they use the same timing. Watch the drop at 3 seconds: the route reveal and bass hit should land together. Listen for effects that overpower the mix.
Step 5: Fix it by frame
Give notes with a time, an object, and a change: “At 5.7 seconds, reduce the glow on the Big Sur card so the text is readable.” Re-render only the affected frames when possible. Then request a 1080x1920 version with camera framing and text adjusted for Reels and Shorts.
P.S. You can access all my AI courses, trainings, prompts library, and $500+ worth of AI tool discounts here.
Perplexity released Decider for Mac, a menu bar app that runs its 4B pplx-decider-flash-v1.2 model locally without an API key. Download 5.8 GB, then label files, route messages, and check screenshots without sending data off-device. Requires Apple silicon, macOS 13 or later, and 16 GB of memory.
Google gave Gemini permission-based access to your Google Maps searches and saved places. Connect Calendar, too, and it can suggest somewhere that fits your schedule, like that restaurant you saved near your last meeting.
OpenAI moved the Decisions API out of beta at $0.10 per million input tokens, with no output-token charges. Updates include prompt caching, US and EU processing, and up to 400 questions per request, double the earlier limit.
🎨 Ideogram: Launched Vector 1.0, its first vector model. Turn a prompt or image into an editable SVG, complete with gradients and typography.
⚖️ Firecrawl: Added a free Government & Legal Index with 30M+ US government pages for ChatGPT and Claude, covering statutes, court opinions, city codes, and permits.
🎬 FLUX Video Upscale: Black Forest Labs upgraded its upscaler to bring clips up to 20 seconds long to 4K, preserving the original audio.
🎙️ Notion: Added an AI Meeting Notes home-screen widget for iOS and Android. Start transcribing in two taps before the conversation gets away from you.
THAT’S IT FOR TODAY
Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!
See you tomorrow :)
- Dr. Alvaro Cintas
✓ Full archive of premium guides with ready-to-use prompts
✓ Structured AI courses (step-by-step, start-to-finish)
✓ Every upcoming premium tutorial
✓ $500+ worth of AI tools