Good morning! Here's what's happening in AI today:
Anthropic restored access to Fable 5, its most powerful consumer model
xAI launched Voice Agent Builder, a no-code platform for human-like voice agents
Notion launched an HTML block for building interactive explainers and prototypes on any page
What are MCPs and why they are AI’s next big unlock
4 new AI tools worth trying today
AI MODEL
Anthropic announced that Fable 5 is back, restoring access to its most powerful consumer model after suspending it earlier this month alongside Claude Mythos 5.
Fable 5 is Anthropic's most capable model for general use, sitting above Opus 4.8 in the lineup and offering the strongest reasoning, coding, and agentic performance available to subscribers outside of the Mythos Preview program.
Access was suspended in June following a US government directive, with Anthropic stating at the time that it aimed to return access as quickly as possible. That return is now confirmed.
The restoration means Pro, Max, Team, and Enterprise subscribers can access Fable 5 again immediately without any additional steps or reactivation required.
Mythos 5 remains unavailable through Claude.ai and is still limited to the Project Glasswing program. Fable 5 is the highest-capability model now available to the general subscriber base.
Fable 5 is the most capable model Anthropic makes available to paying subscribers. Its suspension left a real gap for users doing complex research, long-horizon agentic work, and advanced coding tasks. That gap is now closed. Anyone on a paid Claude plan who downgraded their workflow to Opus 4.8 during the suspension can switch back today.
AI AGENT
xAI introduced Voice Agent Builder, a no-code platform for creating human-like voice agents powered by Grok Voice, available today at $0.05 per minute at x.ai/voice.
Voice Agent Builder lets anyone build a custom voice agent without writing code, using Grok Voice as the underlying model to produce human-like conversation quality across any use case.
At $0.05 per minute, the pricing is transparent and usage-based, giving builders a clear cost structure from day one rather than opaque subscription tiers that make scaling unpredictable.
The platform is live today at x.ai/voice, meaning builders can create, configure, and deploy a voice agent in a single session without waiting for access or going through an enterprise sales process.
Voice agents built on the platform inherit Grok Voice's real-time knowledge, giving them access to current information rather than being limited to a fixed training cutoff the way most voice AI systems are.
Building a voice agent has historically required stitching together a speech-to-text model, an LLM, and a text-to-speech layer - three separate services, three separate APIs, three separate costs. Voice Agent Builder collapses that into one platform with one price. For small teams and solo builders who want to add voice to a product or workflow, this removes the biggest technical barrier that was keeping them out.
AI PRODUCTIVITY
Notion launched an HTML block that lets users build interactive HTML directly on any Notion page, with AI turning existing content into explainers, prototypes, or diagrams that teams can share and tinker with together.
The HTML block lets anyone embed fully interactive elements directly inside a Notion page without leaving the app or managing external hosting, making it possible to build live tools and explainers inside the same document where the work lives.
AI can take existing content on the page and convert it into an interactive format automatically, turning a text explanation into a clickable explainer, a list of steps into a working prototype, or a data set into an interactive diagram.
Finished HTML blocks can be shared with teammates who can modify and build on them together, giving teams a collaborative interface for building lightweight interactive tools inside their existing Notion workspace.
The feature is available now to all Notion users, with no additional configuration or external tool required to embed or share interactive HTML on any page.
Notion has been expanding steadily from a notes and docs tool into something closer to a full work operating system. The HTML block is the most significant step in that direction yet - it means teams can now build and deploy interactive tools without ever leaving the place where their actual work lives. For anyone already using Notion as their primary workspace, this removes one of the last reasons to open a separate tool.
HOW TO AI
What MCPs are and why they are AI’s next big unlock
I connected Claude to my Google Drive two months ago and stopped thinking about file uploads entirely. Claude just reads what it needs, when it needs it. No copy-pasting, no uploading, no context-switching.
That connection runs on MCP. And once you understand what MCP actually is, you start seeing it everywhere.
1. What MCP is
MCP stands for Model Context Protocol. Anthropic built it as an open standard. A universal language that lets any AI connect to any external tool, database, or service.
Think of it as a universal adapter. Before MCP, every AI-to-tool connection needed its own custom wiring. With MCP, one standard handles all of them.
One protocol. Everything connects.
Three pieces, each with a job:
HOST - your AI app. Claude, in this case. It's the brain that decides what it needs.
CLIENT - the connector. Lives inside the host. When Claude needs to read a file from Google Drive, the client is what reaches out.
SERVER - the tool itself. GitHub, Slack, Google Drive, your database. Each one runs its own MCP server that speaks the same language Claude already understands.
Claude asks. The client connects. The server responds. Results come back.
You never see any of this happening. You just ask Claude to pull something and it does.
3. Why it matters right now
Before MCP, connecting an AI to five different tools meant five different custom integrations. Messy, expensive, and usually built for developers only.
With MCP, there are now 2,300+ servers available. GitHub, Slack, Notion, Google Drive, PostgreSQL, Zapier and 9,000+ apps through it. Every major AI company has adopted the standard. The ecosystem is already built.
The tools are there. You just have to turn them on.
4. How to turn it on in Claude today
No terminal required. No config files. No code.
Open Claude Desktop. Go to Settings → Connectors and toggle on the services you want: Google Drive, Gmail, Google Calendar, Slack, Notion. Authenticate with your account. Done.
From that point, Claude can read your Drive, check your inbox, pull your calendar, and post to Slack, all inside the same conversation you're already having.
The first one to turn on: Google Drive. Ask Claude to read a document by name. Watch what changes.
That's MCP. A universal plug that connects Claude to everything you already use.
✓ Full archive of premium guides with ready-to-use prompts
✓ Structured AI courses (step-by-step, start-to-finish)
✓ Every upcoming premium tutorial
Nous Research released Hermes Agent v0.18.0, adding a Mixture of Agents model picker, a /learn command that turns anything into a reusable skill by describing it, a /journey memory graph that shows everything the agent has learned, and background subagents that run without blocking the main chat.
Google announced that Gemini Spark is coming to the Gemini app on macOS, alongside updates to its personal AI agent including custom MCP support, integrations with Canva, Dropbox, Instacart, OpenTable, and Zillow, and connections to Google Tasks and Keep for turning brain-dumps into action items.
Firecrawl launched web-scale /monitor, an always-on search that watches the entire web and pings you or your agent the moment something matching your criteria comes online, expanding the previous single-page monitoring to the full web.
💻 Visual Studio Code: Build and test web apps directly in the browser with VS Code, with pages now able to request camera, location, and device access via per-site approval prompts, plus a new Agents window with guided onboarding tours in preview.
🎬 Higgsfield Shorts Studio: Powered by Gemini Omni Flash, adapts any clip to the formats and pacing that perform on short-form platforms, with editing built to grab attention from the first second. Available on Higgsfield, MCP, and Supercomputer.
🎨 Reve Live Layers: Add text and graphics to any image with fully adjustable layers, giving designers a fast way to create layered compositions without switching to a separate editing tool.
🤖 Manus for Small Business: A dedicated toolkit for solopreneurs and small teams with pre-built connectors, ready-to-run workflows, and live sessions walking you through how to put it all to work.
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