Yesterday's Top Launches: 2 Tools from July 12, 2026

Yesterday brought two interesting additions to the growing ecosystem of AI team assistants, both aiming to embed themselves as true digital colleagues rather than just another chatbot. If you’re feeling the strain of app-switching fatigue or wondering how to integrate new developer tools into your existing workflow without adding more complexity, these launches are worth a look.

Scarlett

Scarlett is positioning itself as a versatile AI co-worker that meets your team where they already are: primarily in Slack, but with a notable twist of also being accessible via iMessage. The core idea is to reduce the friction of bouncing between different applications for different tasks by bringing a powerful AI assistant directly into the communication channels your team uses daily.

Instead of forcing your team to learn a new platform, Scarlett integrates into your existing Slack workspace or even personal iMessage threads, which is a clever move for solopreneurs or small teams who might not live in Slack full-time. It’s designed to function like a real colleague, capable of handling everything from generating a daily company report to triaging customer support tickets. One of the more ambitious features is its “Autopilot” mode. After being trained on a library of business and growth books, Scarlett can be set to autonomously manage areas like marketing or customer support, which could be a game-changer for small teams wearing multiple hats.

From a technical standpoint, Scarlett’s approach to memory is interesting. It uses a hybrid system within a single SQL engine to balance speed and accuracy, aiming to avoid the pitfalls of purely vector-based databases. It also adopts a “Right Model, Right Job” philosophy, letting you direct specific tasks to the AI model best suited for it, whether that’s coding with Sol or design work with Fable. A significant differentiator is its “Use Our Keys” model for API costs. Rather than requiring you to manage subscriptions to various AI services, Scarlett lets you use its keys and simply passes along the cost, which could simplify procurement but is a pricing model some might want to scrutinize.

The target audience is clearly non-technical. The goal is to abstract away the complexity of APIs and models, presenting a simple interface in a familiar messaging app. It’s free to use, which makes it easy to experiment with, but the long-term value will depend on how well its autonomous functions perform in a real-world setting. The iMessage integration is a unique touch, but its effectiveness for serious business operations compared to a dedicated platform like Slack remains to be seen.

Yasmine Works

Yasmine Works takes a more focused approach, embedding itself exclusively within Slack with a compelling premise: transforming individual channels into specialized AI teammates. The problem it tackles is the gap between having access to a powerful AI like Claude and actually having a functional coworker that executes tasks within your workflow. Simply having a chat window open isn’t the same as having an AI that can act on your behalf, and Yasmine aims to bridge that divide.

Its most defining feature is the “one coworker per channel” concept. When you add Yasmine to your #finance channel, it becomes a dedicated finance AI with its own memory and context, separate from the Yasmine instance in your #marketing channel. This isolation is a smart way to prevent context bleed and ensure that the AI’s knowledge and actions remain relevant to the specific domain of each channel. This could be particularly powerful for larger teams where departmental focus is critical.

A major point in Yasmine’s favor is data privacy and control. It operates using your existing Claude subscription, meaning all conversations and data are processed within your own Anthropic account. Your data isn’t pooled or used to train external models. Furthermore, Yasmine runs in “actually isolated” environments, with dedicated compute and storage for each workspace, which is a strong security promise. This architecture also allows its Dev plan to safely execute code.

Perhaps its most user-friendly principle is “She asks first.” With over 500 tool integrations, Yasmine is configured to require explicit approval in the channel before taking any consequential action, like sending an email or updating a database. You maintain granular control, setting permissions for each tool to allow, ask, or block actions. This balances automation with oversight, which many teams will find reassuring.

The use cases are practical: automatically recapping weekly metrics on Monday mornings, drafting emails based on channel discussions, or analyzing data from connected sources. It’s built for teams already committed to the Slack-and-Claude stack who want to add a layer of automation without sacrificing security or control. It’s a freemium product with a 7-day trial, making it accessible for testing.

While both products aim to be AI coworkers, their philosophies differ. Scarlett offers more channels of access and a bolder vision of autonomy, but Yasmine provides stronger built-in safeguards and a more structured, channel-specific approach that might lead to more reliable and secure integrations for business-critical tasks.

Quick Links