Yesterday's Top Launches: 3 Tools from July 12, 2026
New developer tools like Scarlett are embedding AI directly into existing collaboration platforms such as Slack to act as seamless co-workers.
Yesterday brought an interesting crop of tools aimed at integrating artificial intelligence more deeply into our daily work. If you’re tired of juggling multiple apps or wish your team had an extra set of digital hands, these new developer tools might catch your attention. Rather than being standalone applications, they focus on embedding intelligence directly into the environments where collaboration already happens, like Slack or your messaging apps.
Scarlett
Scarlett is designed to act as a genuine AI co-worker, integrating directly into Slack and iMessage. The core idea is to move beyond the typical bot that you have to summon to a separate window. Instead, Scarlett aims to become a seamless part of your team’s conversation, capable of augmenting your capabilities and even running parts of your business on autopilot. It’s built for teams that are drowning in app-switching and want a more unified, intelligent layer over their communication.
What stands out is its focus on reliability. The team claims “she just works,” which is a bold promise in the world of AI assistants, and they back it up with a year of refining their models and infrastructure. The autopilot feature is particularly ambitious; you can set it loose on tasks like marketing or customer support, and it operates based on training from over 50 business and growth books. For solopreneurs or teams that live in iMessage, the dedicated integration is a thoughtful touch.
A clever aspect is the “Use Our Keys” option, letting you access services like HeyGen or XAI through Scarlett without needing separate subscriptions. It passes the cost along directly, which simplifies accounting. It also employs a “right model, right job” strategy, allowing you to choose specialized models for chat, coding, or design tasks. Under the hood, it uses a hybrid approach to memory, combining SQL for structured data with semantic search for relevance, which should make recalling information faster and more context-aware.
Scarlett seems ideal for non-technical founders or teams who want powerful automation without getting bogged down in APIs. If your goal is to automate daily reports, triage customer inquiries, or manage a social media presence, this could be a compelling, free tool to try.
Yasmine Works
Yasmine Works also positions itself as an AI coworker for Slack, but it takes a different architectural approach. Its standout feature is that each Slack channel gets its own dedicated Yasmine instance, complete with its own memory, tools, and context. The Yasmine in your #finance channel won’t know what the Yasmine in #marketing is discussing, which prevents context pollution and makes each one a specialist for its domain.
A significant differentiator is data control. Yasmine runs on your existing Claude subscription from Anthropic. This means your conversations and data are processed through your own account, not a shared pool, which should be a major plus for teams concerned about privacy and data security. Every action that interacts with an external service—like sending an email or updating a CRM—requires explicit approval in the channel. You can configure these permissions on a per-tool basis, offering a solid safety net.
Beyond on-demand tasks, Yasmine can handle scheduled operations, like generating a weekly metrics report every Monday. This transforms it from a reactive assistant into a proactive team member. The promise is to eliminate the manual busywork that often remains even when using powerful AI models.
The freemium model with a 7-day trial makes it easy to test. It’s a strong option for teams already invested in the Claude ecosystem who want granular control over their AI assistant’s actions and a clear separation of contexts between different projects or departments.
Latenode AI Workflow Builder
Latenode tackles a different but related problem: the complexity of building custom automations. Instead of a pre-defined assistant, it’s a builder that lets you create sophisticated AI-powered workflows using natural language. You describe what you want to achieve, and the AI constructs the nodes, logic, and connections for you. This is a boon for anyone who has been intimidated by the technical hurdle of traditional workflow automation tools.
Its intelligent data field mapping is particularly impressive. When connecting different apps, it automatically maps fields if the match is clear. If there’s ambiguity—like two similarly named fields from different sources—it asks you for clarification instead of guessing. This prevents silent errors that can break an automation down the line. The builder also tests each step in real-time as you build, surfacing data type mismatches immediately rather than during a failed run.
The pricing model is based on execution time (CPU time) rather than the number of steps or operations. This means a complex workflow isn’t necessarily more expensive than a simple one; cost is tied to the actual computational resources used. While it lacks full version control, it keeps a conversational log of changes made by the AI, providing some transparency into the iteration process.
With over 5,500 integrations, Latenode is incredibly versatile. Use cases range from automatically enriching leads and adding them to a CRM to sending new Stripe customer summaries to a Slack channel. It’s perfect for individuals or teams that need to automate multi-step processes across various tools but don’t have the time or expertise to code them from scratch.
If any of these tools pique your interest, you can find more details at the links below: