TechCrunch reported on October 27, 2026, that Rohit Patel had left Meta after eight years and launched Ndoc, an AI-native productivity workspace intended to bring presentations, analysis, data, and Python scripts into one environment. The report described a limited alpha and an enterprise pilot, while Ndoc’s site showed conditional email signup on October 28.
How Ndoc organizes work
TechCrunch described three areas: Stage for presentations, dashboards, simulations, and mini apps; Narrative for analysis, numbered citations, generated figures, and explanations; and Model for datasets and Python scripts. The design links a finished presentation to the analysis and data behind it, rather than treating each output as a separate office file.
Patel said users would be able to inspect the Python scripts behind generated work and ask the AI to explain them. He also described memory spanning the user, a document, a team, and a company, with users able to set an audience and tone. Those are Patel’s descriptions of Ndoc’s intended workflow.
The export format Patel described is a single .ndoc.html file that opens in a browser. He said he planned to open-source the format. That would give teams a browser-readable file to share; the open-source release was a plan in the October 27 report.
What access and pricing look like
On October 28, Ndoc’s signup page asked visitors for an email address and said it would grant access if there was room. Otherwise, visitors could join a first-come, first-served waitlist. The page offered a route to request access, with admission conditional on space.
TechCrunch reported that Ndoc was running an enterprise pilot and offering limited alpha access to people on its waitlist. Alpha signup was free; customers were expected to pay based on usage and tokens. The report also said Patel hoped to release a beta in the coming weeks and planned a full launch in 2027.
TechCrunch described Ndoc as a three-person team with no funding at the time of its October 27 report. Its founder, formerly a director at Meta Superintelligence Labs, had spent eight years at Meta working on AI, including reinforcement learning and evaluating AI output.
