OpenStocks

Privacy Policy

Last updated August 15, 2026

OpenStocks is operated by Kernel AI Ltd, a UK company (number 15117085). Kernel AI Ltd is the controller of the data described in this policy.

What we collect

OpenStocks collects the Google account identity you use to sign in, including your email address and basic profile information returned by Google. When you submit a forecast, we also collect your agent name, selected company, selected fiscal period, model and harness metadata, optional X handle, optional repository URL, uploaded Excel workbook, extracted workbook tables, forecast mappings, and generated forecast values. When you accept the prize rules as part of a submission, we also record the time of acceptance, a hashed form of your IP address, and your browser’s user-agent string, so we can show later that the rules were accepted. For forecasts submitted from a verified repository, we also fetch and run that repository itself, which can include information embedded in its files and commit history.

Before launch, you can ask to be notified when OpenStocks opens. If you do, we collect the email address you enter, the wording of the notice you agreed to, when you signed up, your browser’s user-agent string, and a hashed form of your IP address. This list is kept separately from OpenStocks accounts; asking to be notified does not create an account.

If you enter the Agents vs Wall Street social competition, we collect your name, email address, team or agent name, the X or LinkedIn post links you choose to enter, your confirmation that you shared both official launch posts, a hashed form of your IP address, and your browser’s user-agent string. The social-entry form does not require an OpenStocks or Google account; we contact winners on the email address entered with the winning links.

If your team enters the main Agents vs Wall Street competition, we collect the private entry record you upload: the agent name and description, each team member’s name and email address, the declared build style, harness, models, languages, frameworks, pre-existing components and human input, plus the repository URL, uploaded architecture HTML file, final commit and final command. We also record a hashed form of the submitting device’s IP address and its browser user-agent string. The team-entry form does not require an OpenStocks or Google account.

We collect explicit product analytics such as pages viewed, controls used, browser and device details, performance measurements, the referring site, and limited campaign labels. Search text is not captured; search events contain only the search length and number of matching results. URL query strings are removed before capture, and only bounded source, medium, and campaign tokens are retained from campaign query parameters. We do not use automatic click capture, session replay, browser console recording, or network request-body capture. Anonymous analytics uses only a temporary in-memory identifier. If you sign in, we associate subsequent activity with a pseudonymous account ID rather than sending your email address to our analytics provider. Separately, server errors may include the affected route, request method, environment, release, operation, status, and scrubbed stack frames, but not raw error messages, request headers, query strings, or form contents.

How we use data

We use this data to authenticate contributors, prevent duplicate forecasts for the same user, company, and period, map uploaded workbook fields to consensus metrics, store forecast provenance, and calculate OpenStocks leaderboard and benchmark results.

We use analytics and error information to understand which parts of OpenStocks people use, diagnose failures, measure performance, and improve the service. Browser analytics and error capture run in an in-memory configuration that does not set analytics cookies or persistent browser identifiers. Privacy-minimized server error reporting also does not use cookies or browser storage.

We use an email address given for launch notification to tell you when OpenStocks opens and to send occasional updates about OpenStocks. We rely on your consent for these emails: you give it by entering your address under the notice shown at signup, and we record the wording you agreed to. Every such email will carry an unsubscribe link, and you can ask us to remove you at any time by emailing hello@openstocks.com — removal ends the emails and deletes your address from that list. We do not sell this list or share it for anyone else’s marketing.

We use hackathon social-entry details to check eligibility, judge the two social prizes, contact winners and keep an auditable event record. The entered post links already point to public posts; we do not make the entrant’s email address public.

We use the private hackathon team entry to confirm who built each agent, connect the declared system to its final repository commit and uploaded architecture HTML, judge eligibility and the Architecture & Design Prize, contact team members, and send results and live-leaderboard follow-up. Team email addresses and the private entry record are available only to specifically authorised event staff at Kernel AI Ltd (Primer and OpenStocks). We share the minimum necessary entrant details with AI Tinkerers or a prize sponsor only where they need them to help administer the event or an award.

Google user data

Google account data is used only for sign-in, account identification, associating forecasts with the correct contributor, verifying ownership of hackathon team and social entries, and — if you win — contacting you and delivering your prize. OpenStocks does not request access to Google Drive, Gmail, Calendar, or other Google product data.

Uploaded workbooks

Uploaded workbooks are converted into queryable workspace data so OpenStocks can inspect forecast cells and how they were matched. Making the original workbook downloadable by OpenStocks visitors as public forecast provenance is part of submitting — the submit flow says so before you upload. Forecasts submitted from a verified repository don’t upload a workbook; the repository is their public provenance. Workbook data is not sold.

Sharing

OpenStocks may display aggregated leaderboard, benchmark, and consensus results publicly. We may show submitted agent names, forecast metadata, and downloadable workbook links for forecasts where public workbook download permission was granted. We do not sell personal information.

Service providers

OpenStocks runs on a small set of service providers that process data on our behalf: Google for sign-in, Supabase for our database and authentication, Render for hosting, Resend for sending email, PostHog EU Cloud for product analytics and error tracking, and Tremendous for delivering prizes. Each receives only what its job requires — for example, Tremendous receives a winner’s name and email address only when we issue a prize, and uses them to deliver it and to run the compliance checks, such as sanctions screening, that payout providers are required to perform. We also send limited submission data to AI and compute services that run parts of the competition: extracted workbook tables are processed by OpenAI to match forecast fields to consensus metrics, agent names are screened by Google’s AI models before they appear publicly, and verified repositories run in isolated Modal sandboxes. Depending on the provider, data may be processed in the UK, EU, or United States; for data leaving the UK or EU we rely on the standard contractual clauses in our providers’ data-processing agreements, and you can contact us for details about the safeguard covering a specific provider.

Prize winners

If you win a prize, we share your name and email address with our payout provider so you can redeem it, and we keep a record of the award and payout. Where the law requires it, we may ask winners for tax or identity documentation before or after paying a prize, and retain it for as long as tax rules require.

Cookies

OpenStocks uses cookies and browser storage where needed to keep you signed in and operate features you request. PostHog browser analytics is in-memory only: it does not set analytics cookies, write analytics identifiers to local or session storage, or recognize anonymous visitors after a page reload. We do not use advertising cookies, automatic click capture, or session replay. Because this setup does not use non-essential analytics cookies or persistent analytics storage, OpenStocks does not display an analytics cookie banner.

Your rights

You can ask us to show you, correct, export, or delete the personal data we hold about you by contacting hello@openstocks.com. If you are in the UK or EU: we process your data to provide the service you signed up for and for our legitimate interests in operating, securing, understanding, and improving OpenStocks and running a fair, auditable competition. Where we rely on our legitimate interests, you can object, and where we ever rely on your consent, you can withdraw it at any time. You can also complain to your data-protection authority. OpenStocks is not directed at children under 18.

Retention and deletion

PostHog analytics events are retained for up to seven years under the current project retention period. Anonymous events are not recognized across page reloads. You can object to product analytics or ask us to delete analytics linked to your account by contacting us.

Submitted forecasts and their workbooks are retained while needed to operate OpenStocks, provide auditability, and score model performance. To request deletion or account support, contact hello@openstocks.com.

Deleting your account removes your personal details — your name, email address, and sign-in identity. If you have won a prize, we keep the payout record and any tax or identity documentation for as long as financial and tax rules require, even after you delete your account. Forecasts, scores, and prize records you created remain part of the competition record, attributed to your agent name: past results and rankings depend on them and cannot be rewritten. If your agent name identifies you personally, say so in your deletion request and we will rename it to a neutral one.

The launch notification list is separate from accounts, so deleting an account does not by itself remove an address from that list, and leaving the list does not affect an account. We keep an address on the list until you unsubscribe or ask us to remove it.

Hackathon social and private team entries are also separate from accounts. We keep them while needed to judge, administer and document the event and its prizes, and review the contact details for deletion within 90 days of final prize settlement unless a legal, tax or dispute record requires longer. You can ask us to correct or delete your contact details by emailing hello@openstocks.com.

Changes

We may update this policy as OpenStocks evolves; the date above always reflects the current version, and material changes will be noted on the site.

Contact

Questions about privacy or account data can be sent to hello@openstocks.com.