🌍 OpenAI Launches GPT-6.1 Terra

Good Morning! Here's what I have for you in today's newsletter:

OpenAI released GPT-6.1 Terra to read your screen, work across apps, and tackle everyday writing. It’s available through the OpenAI API, with standard pricing 50% lower than GPT-5.6 Terra, making routine computer tasks cheaper to hand off.

Copying data between apps shouldn’t eat the allowance you need for a bigger project. Terra offers a cheaper way to handle forms, replies, and other screen tasks, leaving more room for research and coding with Sol or Astra. Plus and Pro users can check whether it’s available on their accounts.

Aligned shows you what buyers are doing between meetings inside your most important deals then surfaces the risks, openings, and next steps. No more guessing where a deal stands or finding out a champion went quiet too late. Just real visibility, so you can act before it's too late.

Start Using Aligned Free

ElevenLabs introduced the Realtime Skill, giving coding assistants instructions for building with Eleven Realtime. It covers starting a voice agent from scratch and switching an existing browser app to ElevenLabs.

“Add voice to my app” sounds simple until you’re hunting through SDK documentation to explain the setup to your coding assistant. This skill gives it those instructions directly. Whether you’re building something new or moving an existing app, it cuts down on the manual work needed to get started.

Cursor released the Agent SDK for building agents as code, with prompts in Markdown and tools in TypeScript. Everything lives in files alongside the rest of your project.

A tiny prompt edit can fix one problem and quietly create another. Built-in evaluations help you catch that trade-off before users do. Keeping prompts and tools in familiar project files also makes changes easier for teammates to read and review, with a choice between Cursor hosting and your own servers.

Claude turns your notes into a video. ElevenLabs supplies the voice.

Claude Code writes the script, creates captions, and assembles a vertical video from your notes, and ElevenLabs generates the narration. You’ll need Claude Code, an ElevenLabs account with API access, ffmpeg, and a Mac.

Check your tools with claude --version and ffmpeg -version. Install ffmpeg with brew install ffmpeg if needed. Run mkdir -p ~/faceless-short/sources ~/faceless-short/output, then start Claude inside that folder. Ask it to write CLAUDE.md with your topic, tone, banned words, colors, and font. Require a source for every factual claim.

Create an API key with the required permissions and a credit quota. Exit Claude, run read -s ELEVENLABS_API_KEY, paste the key, then run export ELEVENLABS_API_KEY. Restart Claude and choose a voice. Save its ID in CLAUDE.md, keeping the key out of the chat.

Save five to eight facts with links in sources/notes.md. Ask for a 45-60-second script with a strong opening, short beats, and a title for each card. Read it aloud with a timer and check the claims. Cut a weaker beat if it runs long.

Ask Claude to use ElevenLabs’ with-timestamps endpoint, saving output/voice.mp3 and the timing data. Build captions of roughly two to four words from those timings. Listen for pronunciation and pacing before rendering. Each take uses credits, so finish the script first.

Ask for a 1080x1920 ffmpeg render with one card per beat, narration, and burned-in captions. Export three check frames to review text placement. If the subtitles filter is missing, ask Claude to check for drawtext. Watch the finished video on your phone, upload it yourself, and save the tested process as a faceless-short skill.

P.S. You can access all my AI courses, trainings, prompts library, and $500+ worth of AI tool discounts here.

Firecrawl introduced Jev alongside company data in Alexandria. Pull company records, then have Jev classify them for lead scoring, account research, or market maps. In its demo, it sorted 146 startups by customer type in 2.2 seconds.

Cognition introduced the new Devin Security, using Claude Mythos and frontier cybersecurity models to find, prioritize, and fix urgent vulnerabilities. The company says hundreds of millions of dollars in credits will help teams tackle patches quickly.

Perplexity added queued follow-ups to Computer. Enable the setting, and messages sent mid-task wait until the current turn finishes. Edit or delete them while they wait, or send one straight into the running task when you need to change direction.

🧪 Runway: Launched App Studio for building creative apps from a description, using its image, video, and audio models alongside your Brand Kits. No coding required.

🎬 Krea: Launched a Premiere plugin for storyboarding, generating video, and adding sound effects without leaving your editing timeline.

🎵 Lovable: Now generates music for the apps you build. Every track in its demo game was created using Google’s Lyria models.

📣 Fastlane: Takes your website, researches ads in your market, generates thousands of variations, launches campaigns, and tracks the results.

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