Legal
Privacy Policy
What EarlySignal Tracker collects, where information goes, what remains after deletion, and the choices available to you.
Last updated: October 18, 2026
1. Who is responsible and what this covers
Michael Amato, a sole proprietor doing business as EarlySignal, based in New York, USA, is responsible for the information practices described here (“EarlySignal,” “we,” “us,” or “our”). This Policy covers Tracker, Research Lab, Communities, Briefs, and EarlySignal Connect and Market Data processing used to deliver Tracker features. It also covers related account and support communications. It does not describe every practice of an independently operated broker, payment provider, community owner, or linked website.
This Policy is effective October 18, 2026. It describes information practices; reading it or using the Service is not blanket consent to optional processing. Contact michael@earlysignal.co or support@earlysignal.co about this Policy or a privacy request. The Terms of Service address purchases, cancellation, and use of the Service.
2. Information collected and its sources
- Account and authentication: email, name if provided, password hash, verification and reset records, session identifiers, login history, two-factor settings and secret, recovery-code hashes, registration and legal-version records, and timezone, currency, theme, and onboarding preferences.
- Journal and financial records you submit: instruments, symbols, contract details, executions, prices, quantities, fees, dates and times, currencies, account labels, balances, trade plans, notes, strategies, mistake and emotion tags, personal reviews and ratings, screenshots and their filenames, buckets, and other import or editing information.
- Imports and provenance: uploaded file contents during processing and preview, filenames, hashes, mapping metadata, source rows and fields, account and execution identifiers, reconciliation evidence, errors, risk acknowledgments, and import or rollback summaries. Source fields can retain personal or account information present in a file; AI redaction does not remove that information from your stored journal.
- Broker-provided data: authorized account and connection identifiers, brokerage and display names, currency, balance, cash, buying power, holdings, order and execution history, source payloads, checkpoints, and sync status. EarlySignal Connect retains connection tokens and related source records separately from Tracker. Your provider or broker may collect identity or sign-in information directly.
- Research and generated records: questions and clarification answers, study definitions, selected fields and populations, calculated performance and behavioral comparisons, discovery results, pattern states, immutable analytical snapshots and published Briefs, usage counters, and configurations.
- Community and sharing: community names, descriptions and images, member identities, invitation email addresses, delivery status, membership and sharing choices, consent receipts, and relevant activity. We may receive your email from an owner inviting you before you register.
- Billing: plan, price and interval, access grants, subscription and Stripe customer identifiers, payment status, billing-event payloads, transaction and dispute information, usage, and administrative records. Card entry occurs on Stripe-hosted pages; Tracker’s checkout form does not receive full card numbers or card security codes. Stripe events may contain billing contact and payment-method metadata.
- Communications and technical information: support emails, problem reports and their name, email, message and source page; IP address, user-agent, request identifiers, timestamps, errors, access and security events, and selected landing-page button-click events. Hosting and delivery providers also receive the technical information needed to serve requests.
The Service derives metrics, classifications, performance relationships, and other inferences from those records. It does not require precise GPS location, biometrics, a Social Security number, or unrelated health information. Avoid putting unnecessary sensitive information about yourself or others in free text, filenames, images, or uploads. Re-encoding images can remove embedded metadata but cannot remove information visible in an image.
3. Why information is used and who can access it
We use information to register and authenticate accounts; retrieve and reconcile authorized trading history; store and display your records; calculate analytics and currency projections; translate research questions; provide studies, discovery, monitoring and Briefs; manage layouts and sharing; process subscriptions and usage; deliver emails and exports; respond to support; diagnose and improve reliability; prevent abuse and repeat trial use; investigate security or billing issues; and meet legal obligations or resolve disputes.
Authorized operational users can access relevant account information, recent trades and imports, support reports, audit records, and billing or usage details through administration tools or necessary infrastructure access. Your journal is not end-to-end encrypted so that only you can read it. Access is needed to operate the Service and provide support; we do not promise that staff can never view account or trading information. Historical personalization and behavioral comparisons are analytical features, not a professional assessment of your finances or mental health.
4. AI processing and reusable import mappings
Smart Import: depending on the active processing mode and available cached mappings, an AI request may occur before local parsing or as a fallback. It can include structural headers, bounded samples of file rows, categorical values and structural information, symbols, prices, quantities, dates and times, and mapping-validation feedback. The system attempts to filter obvious personal identifiers and sensitive columns; it does not guarantee that every identifying or sensitive value is removed. Do not assume an unfamiliar file is processed entirely on your device or entirely without an AI request.
Research Lab: question translation can send the question and clarification answers, default bucket identifier, available bucket names and identifiers, symbols, setup labels, rule-violation labels, broker-account labels, and supported field options to the AI provider. These inputs are not subject to the CSV sample redaction process. A hashed internal user identifier is also sent for provider safety checks. Avoid including personal information or secrets in questions or labels. Structured outputs are validated before local analysis; the question request is not a transmission of the entire execution database, but its selected context can still be sensitive.
The configured default provider is OpenAI. Both integrations request that API response storage be disabled. This setting is not a promise of zero retention, no provider access, or no prompt caching. OpenAI’s published API policy says inputs and outputs are not used for model training unless its customer opts in, and describes abuse-monitoring retention, ordinarily up to 30 days with exceptions. The applicable provider account settings, agreement, legal obligations, and feature-specific rules govern provider handling. See OpenAI’s data controls. The implemented EarlySignal features use AI inference, not an EarlySignal model-training pipeline; this does not certify a provider’s account-level settings or extend OpenAI’s rules to another endpoint.
Successful import mapping metadata is cached within your account. The Service also builds reusable schema certificates across users from mapping behavior and hashed column/value selectors, with compatibility and successful-import evidence. These shared certificates exclude execution rows, sample trades, filenames, raw source fields, and AI narrative. Related attestations link to contributing users and imports while retained. Hashing reduces direct exposure but does not make all metadata anonymous. This reuse is separate from training a generative model and can continue for a certificate after account-linked evidence is deleted.
6. Service providers and other disclosures
| Provider | Purpose | Information involved |
|---|---|---|
| Vercel | Application hosting and delivery | Requests, network and device metadata, application processing, and operational logs. |
| Neon | Managed PostgreSQL databases | Stored account, trading, research, community, security, support, and billing records. |
| Cloudflare / R2 | Media storage and delivery; domain or network services where used | Uploaded images, archive objects, request metadata, and public community images. |
| Upstash | Request limiting | Counters and hashed account/session or IP-derived identifiers; these hashes are not a guarantee of anonymity. |
| Resend | Account, invitation, and optional Brief email delivery | Recipient addresses and email content, including analytical summaries when Brief email is enabled. |
| Stripe | Checkout, recurring payments, invoices, fraud checks, and billing management | Account email and internal identifier, purchase details, and billing/payment information entered into Stripe; billing events return to EarlySignal. |
| OpenAI | AI file mapping and question translation | The inputs described in Section 4. A different configured AI endpoint, if used, receives the inputs needed for that feature. |
| Trade It and your brokerage | Broker authorization and information retrieval through EarlySignal Connect | Authorization and connection information, account identity, holdings, balances, orders, and executions available through the connection. |
| Market-data and reference-rate sources | Charts, quotes, and reporting conversions | Requested symbols, dates/time ranges, currency pairs, and technical request identifiers. EarlySignal Market Data uses MarketData.app; reporting rates use Frankfurter. |
EarlySignal Connect and EarlySignal Market Data are supporting EarlySignal services, not separate privacy choices merely because they use different databases. Market lookup requests need symbols, dates and related request details; they do not need your complete journal. Providers may use their own subprocessors and may process certain information for independent fraud, compliance, or account purposes. Their own policies apply to direct interactions, including Trade It’s privacy policy and Stripe-hosted payment pages. We do not claim all providers have identical retention or data-use terms.
The current Tracker product has no advertising pixels, cross-site advertising feature, or personal-data-sale integration. Disclosures needed for payment processing, broker connections, and your sharing instructions still send personal information outside Tracker. We may also disclose information when reasonably necessary to comply with law or valid legal process, investigate misuse, protect rights and safety, obtain professional advice, or conduct a business transfer or related due diligence. Any transfer remains subject to applicable privacy obligations; it does not authorize unrestricted new uses of your information.
8. Retention
Retention depends on the purpose, account status, operational need, and applicable recordkeeping or legal obligations. Archiving trades, stopping a share, disconnecting a broker, and canceling a subscription do not by themselves delete stored history. Some analytical snapshots and published Briefs intentionally preserve the evidence available when created instead of being rewritten whenever a trade changes.
- Import previews: server preview plans expire after 15 minutes. Cleanup is scheduled daily to clear expired pending source and normalized data. Completed or failed jobs retain summaries, metadata, and any retained provenance rather than the full pending plan. Imported execution source fields remain with journal records.
- Mappings: user-scoped mapping cleanup targets 90 days without use. Global candidates unused in the relevant update window are eligible for cleanup after 90 days; quarantined or retired revisions after 180 days. Compatible active certificates can remain in use longer.
- General security audit records: maintenance targets removal of IP and user-agent fields after 30 days and events after 365 days. Pending audit-delivery records, infrastructure logs, and separate billing records are distinct; these periods are not a promise covering every log or backup.
- Exports and media: pending archive jobs initially expire three days after request; ready downloads expire 24 hours after completion. Processing and object deletion are asynchronous. Deleted private trade images and archive objects enter a cleanup queue. Public community-image caches and recipients’ copies may outlast removal of the original.
- Billing and trial evidence: separate webhook payloads, billing/admin audit records, and request receipts may survive account deletion and currently have no automated expiry. A fingerprint of the normalized trial email, the original user identifier, and claim date remain to prevent repeat trials; this fingerprint is not anonymous and currently has no automatic expiry.
These scheduled periods describe cleanup rules, not guaranteed deletion to the minute. Backlogs, failed jobs, provider retention, and legal preservation can extend them. Backup and infrastructure-log lifetimes depend on provider configuration; no single fixed deletion deadline for those copies is promised here. Contact us about retained personal information so we can assess necessity, legal requirements, and your rights.
9. Account deletion and its limits
Settings provides permanent Tracker account deletion with identity confirmation. It removes the active account and its related journal, research, preferences, support reports, memberships, and other records tied to that account by deletion rules. General audit records have the account link, network details, and metadata cleared; separate billing, trial, provider, and pending operational records can remain as described above. Billing cancellation must be resolved before the deletion completes. See the billing policy.
Broker connections need separate attention: use Broker Sync’s disconnect control and the provider’s revocation controls before deleting Tracker, or contact support for assistance. Tracker account deletion does not currently guarantee deletion of separately stored EarlySignal Connect records, revocation of its provider tokens, or deletion of your Trade It or brokerage account. Disconnecting clears EarlySignal Connect’s active delegation tokens but retains historical source records and does not delete the provider’s independent account. A privacy request to us can address retained records in our supporting services.
Deletion does not remove someone else’s invitation record containing your email, copies in sent email or a recipient’s possession, an independently imported layout, legally retained financial records, or browser data. Private media objects are removed through asynchronous cleanup. Deleting an owner’s account may leave public community-media objects requiring separate removal; contact us about those images. We cannot promise immediate erasure from every backup, provider, log, or device. These limits do not excuse compliance with a valid statutory deletion request.
10. Your choices and privacy requests
You can update available account preferences, edit or delete journal content, manage connections and sharing, download trading-data exports, and request a queued journal/screenshot archive where storage is configured. Export formats have different coverage and may omit operational, billing, security, or third-party records. For personal information beyond an export, contact us. You can disable optional Brief delivery in Settings and unsubscribe from Brief emails using their link. Essential service and legal notices are separate.
Depending on your residence and whether a law applies, you may have rights to access or know information, correct inaccuracies, obtain a portable copy, request deletion, withdraw consent, limit certain sensitive-data uses, or opt out of sale, targeted advertising, or qualifying profiling. Where available, you may use an authorized agent, appeal a denied request, and complain to the appropriate authority. The Service does not make eligibility decisions about credit, employment, housing, insurance, or similar benefits through its research tools.
Send requests or appeals to michael@earlysignal.co or support@earlysignal.co, identifying your account email, requested action, and state or country so we can identify applicable rights. We may request proportionate information to verify identity or an agent’s authority; never email your password, payment-card number, or recovery codes. We will handle requests within applicable legal deadlines, explain a denial and available appeal process, and not unlawfully discriminate for exercising rights. Retention exceptions and charges, if any, apply only as permitted by law. These rights are not conditioned on buying a plan or first deleting your account.
11. Security and incidents
The application includes password hashing, optional two-factor authentication, session controls, access checks, request validation, and audit logging. Safeguards depend on configuration, operation, and provider services; no service or transmission can be guaranteed secure. We do not claim end-to-end encryption, independent certification, or that backups eliminate data-loss risk. See the Security overview. Report suspected exposure to support@earlysignal.co. We will investigate and provide incident notices to affected people and authorities when required by applicable law; ordinary email terms do not waive any required notice method.
12. Adults and minors
The Service is intended for adults 18 and older. It does not offer a child-account or parental-consent flow, and we do not seek information from minors. An age statement is not a guarantee that every user’s age has been verified. If you believe a minor has supplied information, contact us. When we learn that a user is a minor, we will restrict further processing and address retention or deletion as required by applicable law, including protections for known minors under New York law and for children under 13 under federal law.
13. Location, international users, and changes
EarlySignal is based in the United States. Information can be processed in the United States and other locations where our providers operate; privacy laws may differ from those where you live. We do not promise exclusive U.S. storage or a particular international-transfer certification. If your jurisdiction requires additional safeguards, legal grounds, or notices, those requirements are not waived by using the Service. Contact us about applicable rights or transfer questions.
We will update this Policy and its date when practices change. Material changes will receive a prominent notice in the Service or by email, with advance notice and consent where required before materially different processing begins. A posted update or continued use does not by itself authorize processing that requires separate consent. We will not apply a revision retroactively to defeat existing privacy rights.
14. Contact
Michael Amato, sole proprietor doing business as EarlySignal, New York, USA. Privacy and legal requests: michael@earlysignal.co. Account help and privacy-request routing: support@earlysignal.co.

