Legal

Privacy Policy

Last updated 1 October 2026

This policy explains what information DClaw Forecast collects, why, and the choices you have. In keeping with the product’s principle — nothing invented, nothing hidden — we keep this plain.

What we collect

  • Account information. Your name and email address, handled by our authentication provider (Clerk) when you create an account or sign in.
  • Data you upload. The time-series you upload or paste (dates and values) and the forecast configurations you save. This is stored per account so your work is there when you return.
  • Your Forecast Lab activity. Experiments you save to your history (their settings and results), your Beat-the-model rounds (your guess, your range and the scores), and a daily count of AI answers used, which enforces the daily AI limit.
  • Basic technical data. Standard request and error logs generated by our hosting provider (Vercel) to keep the service running and secure, and anonymous page-speed measurements (Vercel Speed Insights, no cookies, no personal data) so we can see how fast pages load for real visitors.

Cookies and browser storage

  • Sign-in cookies. Clerk sets the cookies that keep you signed in. They are essential: the app can't work without them.
  • Your browser's own storage. The app remembers a few choices on your device (the weather place and measure you last looked at) and keeps a copy of your Lab history there, so it loads quickly and works for a moment offline. It never leaves your browser except as described above.
  • There are no advertising or tracking cookies, and no analytics cookies.

What we do not do

  • We do not sell your data.
  • We do not use your uploaded series to train models for other customers.
  • We do not enrich your data with third-party sources or invent values.
  • We do not send your data to the public data sources. Weather, economic and market series in the Lab are fetched by our server from Open-Meteo and FRED; those requests contain only the place or index being looked up.

How your data is used

Your uploaded series are used only to compute the forecasts, ranges, and explanations you request, and to show your saved work back to you. If you ask the in-app assistant a question, or ask it to read a what-if sentence the built-in rules don't recognise, the relevant forecast numbers and your text are sent to a language-model provider to generate a response.

Sub-processors

  • Vercel — application hosting and logs.
  • Clerk — authentication and account management.
  • Neon — the database that stores your datasets, saved forecasts, Lab experiments and challenge rounds.
  • Vercel AI Gateway — routes assistant requests to the language model.
  • OpenRouter / Anthropic — language-model responses for the assistant (Claude), reached through the Vercel AI Gateway or OpenRouter.

Retention and deletion

You can delete individual datasets and saved forecasts at any time from the workspace, and Lab experiments from the Lab's History. Limits keep storage bounded: up to 50 saved series, 100 saved forecasts, your newest 30 Lab experiments and your newest 500 challenge rounds. Under Account ▾ → Your data & account you can download everything we store for you as one file, or delete your account: that removes all of your data and your sign-in account straight away.

Contact

Questions about this policy? Email hello@smritiraj.online.

DClaw Forecast is an evolving product. This policy is provided in good faith and may be updated; it isn’t legal advice — review it against your own requirements before relying on it commercially.