Analysis
Scuba, operating as Flashback AI Inc., has launched with $34 million raised across two funding rounds, the company said, per SiliconANGLE. Accel led the first round and co-led the second alongside IVP, with BoxGroup, Conviction, Homebrew, and Scribble Ventures also participating. Angel investors include co-founders of Salesforce, Dropbox, and Stripe.
The startup was founded by Noah Desai Weiss, Slack's former chief product officer, and Dolapo Falola, former head of engineering at The Browser Company — the two previously worked together at Google.
What Scuba Actually Does
Scuba calls its product a "context engine": a shared repository where teams store documents, chatbot outputs, and correspondence that both employees and AI agents can draw on. Its Capture feature uses a browser extension and iPhone app to save articles, chat responses, and up to 90 minutes of audio notes; Collections group related files into themed repositories; and Briefing generates a daily AI summary of what changed in a given Collection. An MCP server lets outside AI tools — Scuba names Claude, ChatGPT, Cursor, and Slack — pull from that stored context directly.
The category sits adjacent to enterprise search and knowledge-management incumbents like Glean and Notion AI, both of which have raised large rounds on a similar thesis of unifying a company's scattered knowledge for AI use, as well as newer MCP-native startups building the connective tissue between AI agents and company data.
Scuba hasn't disclosed the individual size of either round, a valuation, or any usage metrics — no customer count, no retention figures, nothing beyond the $34 million total and its founders' pedigree. A founding team with Slack and Browser Company credentials attracts attention and angel checks from well-known operators, but that's a signal about the team, not evidence the product is sticking with early users.
The next real test comes whenever Scuba discloses actual usage numbers — until then, its MCP-server bet on becoming a default context source for other AI tools is the more interesting thread than the funding total itself.
