Yesterday's Top Launches: 5 Tools from August 2, 2026

Yesterday brought another wave of interesting launches, with a particular emphasis on new developer tools and systems that help manage complexity. From wrangling AI budgets to understanding why a mysterious process is hogging your CPU, the releases cover a broad spectrum of practical challenges.

DepthData

If your company is juggling subscriptions to ChatGPT, Claude, Copilot, and a dozen other AI services, DepthData is aiming to be your single source of truth. The core problem it tackles is the financial black hole that AI tool adoption can become. Companies often have no centralized view of their total spend, which teams are actually using the tools, or how many licenses are sitting idle.

What sets DepthData apart is its commitment to verified data. Every figure in its dashboard is explicitly labeled with how it was derived—either ‘MEASURED’ from direct API data or ‘ALLOCATED’ based on usage. It never reads user prompts, focusing instead on metadata like seat assignments and usage patterns, which should ease the minds of security-conscious IT departments. For finance teams trying to reconcile bills or IT managers optimizing software spend, this kind of transparent, audit-ready overview is long overdue. It’s a freemium product, so you can likely get a sense of its value before committing.

Poth Labs

Poth Labs is tackling a different kind of fragmentation: the scattered nature of customer knowledge. Information about your customers is typically locked away in CRMs, support tickets, call recordings, and analytics tools, making it nearly impossible to answer complex questions like, “Why are our most valuable accounts suddenly churning?”

Instead of just being another dashboard, Poth Labs builds a connected model of your customers, linking together data from all these disparate sources. Its key strength is the ability to reason across this entire knowledge base. When it provides an insight, it can show you the specific evidence—a support ticket, a dip in usage data—that led to that conclusion. If the data to answer a question is missing, the system can even trigger adaptive surveys to fill in the gaps. This is a powerful concept for product managers and customer success teams who need to move beyond guesswork and base decisions on a holistic, evidence-based understanding of their users.

witr

Here’s a tool that answers a question every developer and sysadmin has asked in frustration: “Why is this running?” While commands like ps and lsof tell you what is active on your system, witr tells you why. It traces the ancestry of a process, port, or container back to its origin, whether that’s a systemd unit, a cron job, or a shell session.

It’s the context that existing tools lack. You can point it at a PID and it will reconstruct the chain of events that spawned it. It also flags important warnings, like a process running as root or binding to all interfaces. The tool comes with a handy Text User Interface (TUI) for interactive exploration and offers JSON output for scripting. As a free, open-source binary, witr is an instant upgrade to your debugging toolkit, saving you from piecing together clues from multiple commands. It feels like a utility that should have existed years ago.

Screencap

Screencap addresses the problem of institutional knowledge walking out the door when an employee leaves. It’s a tool that records screen activity—clicks, keystrokes, window context—to document how work is actually done. The goal is to turn these messy, real-world workflows into structured datasets, which are gold for training AI models or building automations.

A major and necessary focus here is privacy. The tool is designed to enforce consent and will automatically block recording when sensitive apps like password managers or banking software are active. All data is scrubbed and reviewed on the user’s machine before anything is uploaded. While the potential for creating accurate training data is significant, the idea of constant screen recording will understandably raise eyebrows. The success of such a tool hinges entirely on transparent privacy controls and team-wide trust. It’s open-source and offers a free trial, which is the right approach for a product that needs to prove its security model.

Gemini Robotics 2

Google DeepMind’s Gemini Robotics 2 is the most research-focused launch of the day, aiming to be the “brain” for next-generation robots. This isn’t about simple automation; it’s about giving robots whole-body intelligence, dexterous manipulation, and adaptive reasoning so they can navigate and interact with unpredictable physical environments.

This represents a significant step in bridging advanced AI with physical action. The system allows robots to coordinate their entire structure for complex tasks and even collaborate with other robots. The potential applications are vast, from advanced manufacturing and logistics to assisting in hazardous environments. As a research preview, this is infrastructure for developers and researchers building the future of robotics, rather than a product you can deploy tomorrow. It’s a fascinating look at where the field is headed.

Quick Links

For more details on any of yesterday’s launches, check out the full profiles: