Yesterday's Top Launches: 3 Tools from July 11, 2026
A new tool called Auriko helps developers manage the unpredictable costs of using large language models by acting as an intelligent trading desk for API calls.
Yesterday brought a fresh batch of tools aimed squarely at developers and tech teams looking to streamline their workflows. For anyone building with AI, managing security, or just trying to reclaim a few hours in the day, July 11th offered some compelling new developer tools to consider. The launches covered a surprising range of needs, from optimizing the cost of AI calls to securing hastily built applications and automating daily admin work.
Auriko
If you’re building anything that relies heavily on calling Large Language Models, you’ve likely felt the sting of the monthly API bill. Costs can be unpredictable and quickly spiral, especially when you’re locked into a single provider. Auriko steps in as an intelligent trading desk specifically for LLM calls. Built by people from a quantitative trading background, it treats different LLM providers like trading venues, constantly hunting for arbitrage opportunities to save you money.
The core idea is smart routing. Instead of sending every request to, say, OpenAI by default, Auriko’s engine evaluates a bunch of factors in real-time: token price, latency, reliability, and cache efficiency. It calculates a composite score for each possible path and directs your request to the optimal provider at that moment. A crucial point is that it doesn’t sacrifice quality for cost savings. You can specify the exact model you need, and the system routes within those constraints, ensuring the output meets your standards.
The promised benefit is a significant cost reduction—Auriko’s own benchmarks suggest an average of 30% savings compared to going direct. For teams developing AI agents, content generators, or code assistants, this could translate to substantial operational savings. It’s managed through a unified API, so integration should be straightforward. While the tech stack isn’t detailed, the focus is clearly on the financial and performance logic. It’s a product that makes sense for any team that has moved past prototyping and is now scaling their LLM usage, where every token starts to count.
Perfai Security
The rise of “vibe-coding”—rapidly building applications with AI assistants—has a significant downside: security often becomes an afterthought. You can generate a functional app with user logins and payment processing in minutes, but the complex web of access controls (who can see or do what) is frequently incomplete or flawed. Perfai Security is built to autonomously find and fix these live vulnerabilities in apps created with tools like Replit or Cursor.
Its approach is interesting because it doesn’t just scan static code. It uses a multi-agent system to interact with your live, deployed application. One agent explores every page and API endpoint to understand the app’s structure. A security agent then actively tests access controls, trying to find holes like privilege escalation or data leaks. If it finds vulnerabilities, a third agent communicates with your AI coding tool to guide it in generating the necessary patches.
The promise of fixing security flaws in minutes, without needing deep security expertise, is powerful for small teams and startups. It continuously monitors your app after updates, which is vital because new features can accidentally introduce new security gaps. While it’s a fantastic automated check, it’s worth noting that it focuses specifically on access control. It’s not a replacement for a comprehensive security audit covering all potential vulnerabilities, but for the specific and very common problem of insecure vibe-coded apps, it seems like a highly targeted solution. They’re offering a free tier to try it out, which lowers the barrier to giving it a shot on a staging environment.
Toyo
Switching gears from infrastructure to personal productivity, Toyo is a personal AI assistant that aims to cut down on the administrative drag of the workday. Its unique angle is that it lives in your iMessage and works through voice, acting like an executive assistant that’s just a text or phone call away. The goal is to reduce the time spent context-switching between email, project management tools, and calendars.
Toyo handles a variety of routine tasks. It can triage your inbox overnight, summarizing important emails and even drafting replies in your tone. It helps with meetings by researching attendees beforehand and ensuring action items are captured afterward. You can ask it natural language questions about company information—like current pricing or recent strategy changes—and it will pull answers from connected tools like Notion or Slack. The voice interaction is a standout feature; you can have a scheduled call with Toyo to get your daily plan or just send it a voice note with a task.
The onboarding involves a brief call where Toyo learns about your work and connects to your tools. The idea of an assistant that seamlessly moves between text and voice is appealing, especially for busy founders or professionals who are constantly on the move. However, its effectiveness will almost entirely depend on the quality of its integrations and its understanding of context. If it can accurately capture the nuance of a project update or draft a coherent email, it could be a huge time-saver. If it frequently misunderstands, it might create more work. It’s a high-potential tool for anyone feeling overwhelmed by the overhead of modern work communication.
Yesterday’s launches show a clear trend towards practical AI tools that solve immediate, costly problems. Whether it’s saving money on API calls, securing apps built at speed, or automating personal busywork, these products are focused on tangible efficiency gains.
Quick Links