Greetings. Let's dive into what's happening with AI tools and features right now.
Desktop Agents Are Having a Moment
What's been particularly noticeable this week is how Anthropic is pushing to transform its desktop app into something closer to OpenClaw. They already have scheduled tasks, scheduled cloud tasks, and the ability to talk to Claude Desktop from mobile. They're also experimenting with Telegram integration, meaning you could have Claude running on your laptop and communicate with it through Telegram in exactly the same way you do with OpenClaw. It can execute any task on your laptop. A nice and quite obvious direction.
At the same time, other companies are following Anthropic's lead here. Manus launched its own desktop app. xAI is working on something called Grok Computer, which, based on the name, sounds more like they are following Perplexity, though it's too early to judge. The bottom line: desktop apps that can operate your laptop and handle whatever you throw at them are gaining serious traction.
The Enterprise Race
It was also interesting to observe how these companies are planning to bring their tools to enterprise customers. This is another race entirely, and it will have a huge impact on businesses; whoever wins there will hold a significant advantage. Perplexity is pushing Comet for enterprise, Anthropic is rolling out new enterprise tools with Claude, and the competition is heating up.
Google: Lots of Expectations, Not Much to Show Yet
When it comes to Google, there are, as always, a lot of expectations for them to release something. However, things aren't quite there yet. They're working on Skills, Projects, and something called "Build with Gemini" for Gemini Business, which could be interesting. A lot of consolidation is happening behind the scenes.
Smaller Model Releases
On the model release side, we got a wave of smaller models: GPT-5.4 Mini and Nano, Mistral Small 4, a more optimized Composer 2 from Cursor, and the Minimax M2.7 upgrade. Plenty of things to test if you're in the mood.
And now, let's break it all down into granular pieces!
Newsletter Guide
Here's your guide on how to make the most out of this newsletter:
- 🤳 Gear up with your mobile phone and desktop devices.
- 📩 Scan through the newsletter, picking out the apps that pique your interest.
- 🔥 Keep an eye on the "hot" emoji to identify the hottest user experiences.
- 👀 Dive deeper into our TestingCatalog posts to discover how to put these apps to the test.
- 📲 Roll up your sleeves and try them out yourself!
This newsletter was sponsored by
CodeRabbit

CodeRabbit launches AI Planner for team project planning – CodeRabbit launches Planner, a collaborative planning layer for dev teams to align on implementation before any coding agent writes a line of code.
HypeScribe

HypeScribe launches AI transcription and summary tool – HypeScribe converts audio, video, and social media links into accurate transcripts, summaries, and action items in seconds.
Perceptis AI

Perceptis launches AI platform for generating data-driven slides – Perceptis AI delivers MBB-quality presentations grounded in your own data, with editable PPTX output and SOC-2 security built in from day one.
Anthropic

🔥 Anthropic prepares Projects for Claude Cowork on desktop – Anthropic is developing a project layer for Claude Cowork on Desktop, allowing users to organize work in local folders and set scheduled tasks by project.
- 🔥 Claude Code now has scheduled cloud tasks executed on remote machine with MCP access.
- Projects for Claude Cowork officially announced with dedicated local folders and project-specific scheduled tasks.
- 🔥 Claude Code channels released in research preview for Telegram and Discord.
- 🔥 Cowork Dispatch rolled out to all Claude Pro users allowing mobile trigger of desktop tasks.
- Anthropic rolling out Dispatch feature for Claude Cowork mobile-to-desktop communication.
- Anthropic hosting "Code with Claude" developer conferences in SF, London, and Tokyo.
Cursor

Cursor launches Composer 2 AI coding model for all users – What's new? Composer 2 is an AI coding model available to all users on Cursor with an early alpha interface; it has standard and fast pricing variants with proven code benchmarks.
- Cursor debuts CursorBench-3 to evaluate coding agents.
- Cursor released new "Glass" unified workspace for agents, repositories, and cloud tasks.
Fal

Google tests Skills and "Build with Gemini" tools for Gemini Business – Google Gemini Enterprise is testing "Build with Gemini" for app prototyping and a new "Skills" section, both expected to debut around Google I/O 2026.

Google expands Gemini API with unified tool support for developers – What's new? Google adds Gemini API tooling that joins built-in tools with custom functions in one call; cross-tool context circulation and unique response IDs share outputs;

Google Cloud previews G4 VMs using NVIDIA Blackwell GPUs – Google Cloud introduces fractional NVIDIA Blackwell GPU rentals in G4 VMs, offering flexible, cost-efficient options for AI, rendering, and enterprise workloads.
- Google's AI Studio roadmap for next weeks including agents, G1 support, and planning mode.
- Google testing dedicated Gemini desktop app.
- Google released new full-stack vibe coding experience for AI Studio with Firebase support.
- Google exploring Google Play Books as source for NotebookLM.
- Google released new Vibe Design experience for Stitch with AI-Native Canvas.
- Projects feature coming to Gemini way before 6 months.
- Google rolling out Personal Intelligence in Gemini app and Chrome.
- Google rolling out Cinematic Video Overviews in NotebookLM to Pro accounts.
Lovable

Lovable launches AI platform for document, data, and app creation – What's new? Lovable AI platform analyzes data, creates docs, images, videos and turns files to apps; update runs code for data tasks, converting spreadsheets and pdfs to apps;
Manus AI

Manus launches My Computer with local AI agent on desktop – What's new? My Computer lets the AI agent manage files, execute command-line tasks and control apps on macOS and Windows with local GPUs and resources;

Manus AI prepares WhatsApp Business integration for its Agent – Manus AI is set to bring WhatsApp integration after its Telegram debut, requiring WhatsApp Business for linking.
Meta
- Meta working on Citation controls for Meta AI to configure web and social search citations.
- Meta working on AI Detector feature for Meta AI.
- Nebius signed multi-year AI infrastructure agreement with Meta up to $27B.
Microsoft

Microsoft tests Portraits in Voice Mode for Copilot, Sage, and Pax – Microsoft Copilot is testing new voice avatars Sage and Pax plus a OneDrive sync option, with public rollout expected later this year.
- ICYMI: Microsoft begins phased US rollout of Copilot Health.
- Microsoft released MAI-Image-2 image generation model and MAI Playground.
Midjourney
MiniMax

MiniMax launches M2.7 model on MiniMax Agent and APIs – What's new? M2.7 model uses agent harnesses and reinforcement learning for self-updating skills; available via MiniMax Agent and MiniMax API Platform for professionals;
Mistral AI

Mistral releases Mistral Small 4 model under Apache 2.0 licence – What's new? Mistral small 4 is an open source ai model with 119b parameters, mixture-of-experts and multimodal text-image support via several channels; it cuts latency 40 percent and triples throughput vs small 3;

Mistral and Nvidia join forces to co-develop open-source models – Mistral AI joins Nvidia's Nemotron Coalition as a founding member, releasing Mistral Small 4 model for broad use in open-source and enterprise AI projects.

Mistral AI launches Forge for building AI enterprise models – What's new? Mistral AI launched Forge for enterprises to train models with internal data; it offers multimodal input and agent-first tuning;
Okara
OpenAI

🔥 OpenAI launches GPT-5.4 mini and GPT-5.4 nano on APIs – OpenAI introduces GPT-5.4 mini for faster, cost-efficient coding copilots, automation, and agent systems, now available in API, Codex, and ChatGPT. GPT-5.4 nano is now available via the API as well.
- ICYMI: OpenAI adds interactive math & science tools to ChatGPT.
- Sam Altman co-hosting World Lift Off event on April 17 amid rumors of OpenAI Super App.
- OpenAI planning to launch Super App unifying ChatGPT, Codex, and Atlas.
- OpenAI acquired Astral to enhance Codex ecosystem with open-source Python dev tools.
- OpenAI updated ChatGPT model selector on web and mobile for Instant/Thinking modes.
- OpenAI in talks with TPG, Bain, Brookfield, and Advent for enterprise AI deployment.
Perplexity

Perplexity tests Market Research agent for Perplexity Computer – Perplexity is developing a Market Research section powered by Perplexity Computer, offering research workflows and premium data.

🔥 Perplexity launches Perplexity Health agent in US – Perplexity Health is set to offer users secure access to health data providers, personal dashboards, and specialized AI agents within Perplexity.

Perplexity launches Comet Enterprise with advanced security – Perplexity launches Comet Enterprise, a managed AI browser for companies with admin controls, policy support, enterprise security, and data compliance features.
- Perplexity Comet browser now available on iOS with agentic capacity and voice mode.
- Perplexity Computer now available on Android.
xAI
- Early trace of Grok Computer with browsable "Files" side panel discovered.
- Voice mode on Grok rolled out on web and Android for X app.
- xAI working on Yolo mode and Connectors for upcoming Grok Build.
- Grok Text-to-Speech API released with multiple voices.
- X AI recommendation algorithm major update rolling out next week and will be open-sourced.
Xiaomi
Z.ai
Happy testing! 🤖
P.S. Some ref links
- Flowith AI (3k extra credits)
- Ideogram (100 extra credits)
- Runable (1000 extra credits)
- Manus AI (500 extra credits)
- NordVPN (1 month free)
P.S.S.
If you use Telegram, don't forget to join us there too! It is the fastest place to learn about new AI scoops 🗞️
P.S.S.S
If you want to promote your AI tool in this newsletter, please check our Stan Store for Advertisers.