byline.fyi Privacy Notice
Effective date: 8 July 2026
Last updated: 17 September 2026
This is the interim privacy notice for byline.fyi. It describes how the live product handles data today. It is not a substitute for formal legal review.
Who operates byline.fyi
The operator identity below is a candidate placeholder for operator and counsel confirmation, not approved legal wording. These fields must be confirmed before publication.
- Legal name: [PLACEHOLDER: operator legal name — to be confirmed]
- Legal form: [PLACEHOLDER: operator legal form — to be confirmed]
- Country: [PLACEHOLDER: operator country — to be confirmed]
- Business address: [PLACEHOLDER: business address to publish — to be confirmed]
Questions or privacy requests: [email protected].
What byline does
byline reads content you connect or forward, builds a creator profile and memory for your workspace, and makes that memory available to the AI tools you choose to connect. Your workspace data is scoped to your workspace.
Data we collect
Account and workspace data
- Name, email address, sign-in status, sessions, and optional Google sign-in identifiers.
- Workspace name, members, roles, invitations, onboarding state, subscription status, plan limits, and audit events.
Content and source data
- Published work from connected sources such as websites, Instagram, and YouTube.
- Private notes or drafts you add or forward by email.
- Audience comments and engagement metrics where a connected platform exposes them.
- Derived data generated to provide the service, including summaries, chunks, transcripts, embeddings, memories, profile facets, and insights.
- Source credentials, such as platform access tokens, stored so byline can sync the sources you connect.
Billing data
Stripe handles payment details. byline stores Stripe customer, subscription, plan, and invoice identifiers needed to run billing and entitlements. byline does not store full card numbers.
Logs, security, and usage data
byline keeps operational logs, audit events, rate-limit counters, COGS counters, request metadata, and similar records needed to run, secure, and debug the service.
Advertising, analytics, cookies, and Meta
byline uses strictly necessary cookies for sign-in, security, and normal product operation.
byline sets first-party analytics and advertising cookies when you visit. byline's
own attribution cookie (byline_attr) and, when DataFast is enabled, the DataFast
visitor cookie (datafast_visitor_id) are set directly by the site on load, so
they are not gated by the tag manager. When the operator loads Google Tag Manager
and configures consent management there (Google Tag Manager Consent Mode), the
choices you make in those controls govern the additional analytics and advertising
tags the tag manager itself fires.
Separately, when configured by the operator, byline sends server-side conversion events to Meta's Conversions API and payment events to DataFast. These server-side events do not set browser cookies and they still run regardless of your browser cookie choices. The events may include:
- for Meta events, a hashed email address, hashed name, and hashed byline user id;
- for DataFast events, the visitor id plus a plain email address, name, and account identifiers;
- Meta click identifiers such as
_fbpand_fbc, if present; - IP address and user-agent for the Meta sign-up event only;
- payment value and currency for
Purchaseand DataFast payment events.
Paid Subscribe and Purchase events do not replay stored sign-up IP address or
user-agent.
byline also runs a daily retention sweep that nulls stored sign-up IP address and
user-agent after roughly 180 days while keeping non-PII click and campaign ids.
Meta receives this data for advertising conversion measurement and may process it under its own business tools terms. This is a real external processing path, not a browser-only pixel.
Who receives data
byline uses a small set of service providers:
| Provider | Purpose | Data shared |
|---|---|---|
| OpenRouter | Routing, privacy controls, and usage accounting for DeepSeek text and image processing | Content, notes, comments, images, and derived prompts needed for the requested processing |
| DeepInfra | DeepSeek V4.1 Flash model inference through OpenRouter's zero-data-retention route | Content, notes, comments, images, and derived prompts needed for the requested processing |
| Google AI Studio / Gemini | Video understanding and embeddings | Video URLs or uploaded media, text content, and derived text needed for embeddings or analysis |
| Google Sign-in | Optional sign-in | Google account identifier and basic profile if you choose Google sign-in |
| Google Tag Manager | Optional browser analytics | Browser analytics and tag data, subject to your consent choices in the tag manager, and only when configured |
| DataFast | Product analytics and revenue measurement | Anonymous visitor id, payment amount and currency, and account identifiers such as email needed to measure conversions and subscriptions |
| Meta Platforms, Conversions API | Server-side advertising conversion measurement | Hashed account identifiers, click ids, and the sign-up IP address and user-agent described above |
| Meta Platforms, Instagram Graph API | Instagram source connector | Instagram content, comments, metrics, and API requests you authorize |
| Postmark | Transactional and inbound email | Sign-in emails, support emails, and forwarded inbound messages |
| Stripe | Payments and subscriptions | Billing identifiers, payment metadata, and subscription state |
| Cloudflare | DNS, secure ingress, and edge delivery | Request metadata such as IP address and user-agent |
| Hetzner Object Storage | Disaster-recovery database backups | Backup copies of application data, retained under the infrastructure backup policy |
The application and primary Postgres database run on operator-managed infrastructure. The infrastructure repository owns the current backup configuration; this notice names Hetzner because database backups are an external processing path even though the live database is self-hosted.
AI providers
Content is sent to AI providers only to provide the product features you request.
Retention
- Workspace content and derived data are kept while the workspace exists, unless you delete the content or remove the source.
- Disconnecting a source deletes that source's stored content and audience data.
- Billing records are kept as needed for accounting, tax, fraud, and dispute handling.
- Audit logs are retained for security and accountability and may outlive the workspace or user they refer to.
- Disaster-recovery database backups are retained for about 30 days.
- Advertising attribution IP address and user-agent are nulled after roughly 180 days. Click ids and campaign parameters are kept with the user account unless the account is deleted.
Your choices and requests
The first-party byline_attr attribution cookie and, when DataFast is enabled,
the DataFast visitor cookie are set on load and are not governed by the tag
manager's consent controls; you can block or clear them through your browser
settings. The tag manager's consent controls govern only the additional analytics
and advertising tags the tag manager itself fires. You can disconnect sources or
delete individual content inside the product. For access, correction, export,
deletion, objection, or other privacy requests, email
[email protected].
Security
byline uses workspace scoping, signed webhooks, rate limits, COGS controls, central log redaction, and audit logging. No internet service is risk-free. Report suspected security issues to [email protected].
Changes
This notice may change as the product and formal legal review progress. The live
version at /privacy is the current public notice.