Ndoc challenges Microsoft Office and Google Workspace — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Rohit Patel is departing Meta after eight years of AI work to launch Ndoc, the AI-native workspace he is pitching as a replacement for Microsoft Office or Google Workspace; he previously served as a director at Meta Superintelligence Labs.
- Ndoc combines three views in one layout: Stage holds slide decks, dashboards, simulations and mini apps; Narrative contains analyses, numbered citations and generated figures; Model houses datasets and Python scripts.
- Ndoc surfaces the Python scripts used to create documents, letting users ask its chat to explain a formula or code line and verify the work; the design reflects Patel’s background in reinforcement learning and AI-output evaluation.
- Ndoc is built and fine-tuned with a mixture of open-source models and, according to Patel, costs considerably less to run than alternatives built on frontier AI models; early trial users focused instead on seeing all their work end to end in one place.
- Ndoc plans to export each document as a single
.ndoc.htmlfile that opens in any browser and to open-source the format, letting developers build compatible apps without locking users into Ndoc, Microsoft or Google. - Ndoc is running an enterprise pilot and a limited wait-list alpha as of Tuesday, with a beta planned for coming weeks and a full launch next year; its three-person team has no funding, signup is free and customers pay for usage and tokens.
- Ndoc includes connectors that import data from tools customers already use, while Patel says he understands how hard it is to replace entrenched workflows and says OpenAI’s Space, launched inside ChatGPT last month, could be a threat to Ndoc.
Why it matters: With a three-person, unfunded team, Ndoc is pressing its enterprise case before a beta planned for coming weeks, against Microsoft Office, Google Workspace and OpenAI’s Space. For enterprise users, the concrete distinction is source-level transparency—inspectable Python and verifiable formulas—while connectors let them import existing data into one end-to-end workspace.
Ask SkimNews
