Former Slack executives launch Scuba with $34m for an AI context engine

Former Slack executives Noah Desai Weiss and Dolapo Falola have launched Scuba, a startup building a context engine for people and AI agents. The company has raised $34 million across pre-seed and seed rounds, according to Upstarts Media, and aims to make company knowledge easier to bring into the tools where work happens.

Watch Desk analysis

What happened

Scuba lets users collect web material into shareable collections and connect those collections to tools including ChatGPT, Claude and Slack through MCP connectors. The idea is to help workers and agents align around company operations and goals, rather than leaving useful context scattered across tabs and chat threads.

Upstarts Media reports that Accel, IVP and technology founders backed the funding. The report identifies the founders as former Slack executives; it does not detail the company’s pricing or say when the product will be broadly available.

Why it matters

AI agents are only as useful as the information they can access and interpret. Scuba’s pitch is to give them a shared, organised context layer, while also giving people a way to gather and pass along relevant material. That puts the startup in a busy part of the AI market: the connective tissue between company information and the tools that act on it.

The practical test will be whether the collections stay useful as information changes, and whether connecting an agent to company context saves work without creating a new chore of keeping the context tidy. A fresh bucket for all your tabs is not automatically a breakthrough; the trick is making it earn its place.

Our read

Scuba has a clear problem to solve and a sizeable funding cushion for an early-stage company. Its appeal will depend less on the phrase “context engine” than on whether teams can reliably find, share and use the right information across their existing tools. For now, it is a launch and a product proposition, not evidence that the context problem is solved.

What to watch

• Which tools and MCP connectors Scuba supports as the product develops.
• How it handles permissions and changes to information shared with agents.
• Whether customers can show that the context layer improves real work, rather than adding another place to manage it.

Discussion spark: Should companies build a shared context layer for AI agents, or keep business knowledge inside the tools and systems that already own it?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.