Yesterday's Top Launches: 5 Tools from July 21, 2026

Yesterday saw another interesting batch of product launches, particularly for developers and creators looking for new tools. From AI infrastructure to career management, the releases covered a wide spectrum of needs. If you’ve ever felt overwhelmed by online pricing, wanted to run LLMs faster on your Mac, or wished you had a personal agent handling your job search, July 21st brought some noteworthy solutions.

How Old Do I Look

We’ve all been curious. How Old Do I Look is a straightforward, free web tool that uses AI to estimate a person’s apparent age from a photo. It’s built for speed and privacy, processing your image directly in the browser without storing it. The tech stack is modern and lean—Next.js, React, and TypeScript with Tailwind CSS—which suggests a fast, clean user experience. This isn’t a complex analytics platform; it’s a simple utility that does one thing well. It’s the kind of tool you might use for a bit of fun or to quickly test an AI model’s perception, perfect for a quick demo or a lightweight integration into a larger project.

Spycost

Online shopping can feel like a psychological game, with “limited time offers” and “huge discounts” that might not be what they seem. Spycost tackles this head-on by acting as a price-tracking watchdog. You add a product URL, and it begins building a history of its price fluctuations. The real value is in its transparency: it shows you the actual price trend over time, helping you spot if that 50% off banner is masking a recent price hike.

One thoughtful aspect is its approach to regional and session-based pricing. Instead of giving you a potentially irrelevant average, it aims to provide context-specific data based on your own shopping conditions. The vision is to eventually let you track entire product categories, not just single items. It’s a free web tool for now, with plans for a browser extension. For anyone who hates feeling like they’ve been duped by a clever marketing tactic, Spycost offers a dose of data-driven clarity.

Kobbe

As privacy regulations tighten and users become more aware of tracking, the demand for ethical analytics is growing. Kobbe enters this space as a privacy-focused alternative to giants like Google Analytics. It’s built to be cookie-less by default, meaning it collects essential website data—traffic, popular pages, referral sources—without invasive tracking.

Where Kobbe gets interesting is in its specific integrations. It connects with payment processors like Stripe and Paddle to show you which marketing efforts are actually driving revenue, which is gold for SaaS founders and online businesses. It also breaks out traffic from AI agents like ChatGPT and Claude, a smart move as AI-driven discovery becomes more common. You can embed analytics widgets directly into your own site and set up custom event tracking. It’s a freemium tool with a 15-day trial, squarely aimed at developers, bloggers, and indie hackers who need actionable insights without compromising their users’ privacy.

BaseRT

For developers working with AI on Apple hardware, BaseRT is a significant release. It’s a high-performance LLM runtime engineered specifically for Apple Silicon. The key claim is speed: it’s reportedly up to 6.4x faster than llama.cpp in prompt processing by leveraging the Metal API and tensor cores in M-series chips. This isn’t a minor optimization; it’s a specialized tool designed to push local inference to its limits.

The appeal is clear: run models like Llama 3 or Gemma locally on your Mac, which means total privacy and no per-token costs. The installation is a single command, making it accessible. The open-source nature invites community contribution and trust. This is a tool for building always-on AI agents, real-time voice applications, or any project where low latency and data sovereignty are critical. If you’re developing AI applications on a Mac, BaseRT is worth a very close look.

Double

What if you had an agent constantly working to advance your career? Double is an ambitious attempt to be exactly that—a personal AI career agent. It addresses a common frustration: talented people often miss great opportunities because they’re bad at networking, applying, or personal branding. Double proactively searches for and applies to jobs it deems suitable for you, manages your professional network, and works to make recruiters seek you out.

It uses a custom memory layer to learn from its actions, and you can ask it to explain why it recommended a particular role. The vision is to replicate the kind of career management that celebrities have, but for everyone. It’s a free web service for now, with plans to integrate with Google Calendar and Mail. The concept is powerful, though its effectiveness will depend heavily on the AI’s judgment and the quality of its outreach. For someone feeling stagnant or invisible in their career, it represents a fascinating, hands-off approach to professional growth.

Quick Links

For a deeper dive into any of yesterday’s launches, check out the full details here: