Privacy Policy

Last updated: October 29, 2026

1. Introduction

Letter ("we," "us," or "our") is an AI-powered content editing and publishing platform operated by The Letter Company. This Privacy Policy explains how we collect, use, disclose, and protect your information when you use our service at letterstory.com.

2. Information We Collect

Account Information

When you create an account, we collect your email address, full name, and an optional avatar image. Authentication is handled through Supabase Auth, using either an email and password or Sign in with Google (OAuth).

Organization Data

If you create or join an organization, we store the organization name, domain, branding preferences (colors, logo, writing style, tone, and custom guidelines), member roles, and your role within the organization.

Content

We store the content you create, import, or upload — including edits, posts, stories, and collections — along with their full text, titles, metadata, and version history. Version history includes content snapshots and change logs to support versioning.

AI-Generated Content

When you use our AI features (editing suggestions, SEO analysis, content grading, table generation, Venn diagram generation, image generation, and flow processing), your content is sent to third-party AI providers for processing. Text is processed by the Anthropic Claude API; AI image generation is processed by our configured image provider (OpenAI or Bloom). The AI-generated results (suggestions, visualizations, images, and graded outputs) are stored in our database.

Connected Integrations

If you connect a third-party content platform — Notion, Contentful, Webflow, Sanity, Framer, or Google Docs — we store the identifiers needed to sync with it (such as workspace, site, space, or project IDs and names) and the credentials for that connection (an OAuth access token or API key) encrypted at rest. We also log sync operations (import/export), including item titles and sync status. We access a connected platform only to search, import, and publish content as you direct.

Code Repositories (GitHub and GitLab)

You can connect a GitHub or GitLab repository so Letter can make changes to your site (for example, adding structured data, an llms.txt file, or page copy you ask for), publish posts to it as files, import existing posts from it, or watch its pull requests and releases for things worth writing about. We access a repository only when you or your organization's members (or an agent using your API key) ask for one of these, and only the repositories you grant.

How we connect. On GitHub, you install the Letter GitHub App and choose which repositories it can reach. We store only the installation ID; each time we touch a repository we request a token from GitHub that is limited to that one repository, lasts at most one hour, and carries only the permissions the step needs (reading contents, reading checks, or writing a branch and opening a pull request). You may instead provide a GitHub personal access token. On GitLab, you sign in through GitLab OAuth (scopes api and read_user, so we know which account connected) or paste a project or group access token. Personal access tokens and GitLab credentials are encrypted at rest.

What happens to your code. When you ask for a change, we download a snapshot of the repository at one commit into a temporary, isolated build environment created for that change alone. An AI coding agent (Anthropic's Claude, or a model from OpenAI) reads the code it needs and makes the change there, so portions of your code are sent to that AI provider for processing. The build environment never receives your GitHub or GitLab credentials. The change is opened as a pull or merge request on a new branch for you to review; Letter merges it only when someone in your organization asks it to, or when your organization has turned auto-merge on. Where a change can be previewed, the build environment serves a preview of your site at an unlisted link for up to 72 hours, using any preview environment variables you provide (stored encrypted and never shown back). The build environment is then deleted, along with its copy of your code.

What we keep. We do not keep a copy of your repository. We store the instructions for each change, its status, a summary of what the agent did, the list of files it changed and the pull request link, and a log of the agent's steps (which can include excerpts of files it read) so you can see what happened and we can diagnose failures. The log of the agent's steps, including those excerpts, is also sent to Lemma (see Third-Party Services). For posts, we store the posts we import from or publish to your repository. For repository watches, we store the titles and descriptions of the pull requests and releases we read.

You can remove access at any time by disconnecting GitHub or GitLab in Letter, by uninstalling the Letter GitHub App or removing repositories from it in your GitHub settings, or by revoking the token or OAuth authorization in GitLab.

Google User Data (Gmail)

If you connect a Gmail or Google Workspace mailbox to send email through Letter, we ask you to approve two Gmail API scopes — https://www.googleapis.com/auth/gmail.send (send email on your behalf) and https://www.googleapis.com/auth/gmail.readonly (read your messages) — plus openid and email, so we know which mailbox you connected. We do not request full mailbox access. We use that access only to (1) send the email sequences you create from your mailbox, (2) read your mailbox to detect replies, bounces, and opt-outs to messages we sent so a sequence stops, and (3) show you a read-only list of that mailbox's sent and received messages inside Letter. We never delete, move, or modify your messages. We store only your encrypted refresh token and the outcome of each reply check (replied, opted out, bounced, or none); we do not store the content of your mail. You can disconnect the mailbox in Letter at any time, or revoke access from your Google Account permissions page.

Letter's use and transfer of information received from Google APIs to any other app will adhere to the Google API Services User Data Policy, including the Limited Use requirements. The use of raw or derived user data received from Workspace APIs will adhere to the Google User Data Policy, including the Limited Use requirements. We do not use Google user data to develop, improve, or train generalized AI or machine-learning models, we do not sell it, and we do not use it for advertising. Content from a mailbox you connect is not sent to Anthropic, OpenAI, or any other AI provider.

Site Deployments

If you deploy a hosted blog or site, we store the deployment configuration and, through our hosting provider (Vercel), the domains you connect. If you register a new domain through the service, we collect the domain name and pass the registration details required to purchase it to the domain registrar.

API Keys

If you create an API key for programmatic or MCP access, we store a hashed version of the key (never the key itself in plaintext), its scoped capabilities, and a log of when it is used.

Uploaded Files

We store files you upload, including organization logos (up to 2 MB) and custom fonts (up to 5 MB), in Supabase Storage.

Usage Analytics

We use PostHog to collect product-usage analytics — events such as sign-ups, logins, and feature interactions, along with basic device and browser information — to understand how the service is used and to improve it.

Website Visitor Identification

We use RB2B to identify business visitors to our website. RB2B matches a visitor to a professional profile — typically a name, work email address, company, and LinkedIn profile — using data it has collected independently, and tells us which pages that visitor viewed. We use this to contact companies that may be a fit for Letter. Identification is limited to visitors in the United States, and it runs only on our public pages — the marketing page, knowledge base, documentation, prospect reports, and sign-in pages. It never runs inside your signed-in workspace, and never on invite, review, guest-interview, investor, or shared links. To opt out, see RB2B's opt-out page.

3. How We Use Your Information

  • Provide the service — process your content through AI enhancement pipelines, generate visualizations and images, and manage your content.
  • AI processing — send content to Anthropic's Claude API and to our image-generation provider to produce editing suggestions, SEO improvements, grades, data tables, Venn diagrams, and images, and send repository content to Anthropic or OpenAI models when you ask for a code change.
  • Publishing & sync — import content from and publish content to the third-party platforms you connect.
  • Code changes — read the repositories you connect, make the changes you ask for in an isolated build environment, and open pull or merge requests for your review.
  • Deployments — provision and host the blogs or sites you deploy, and register the domains you connect.
  • Programmatic access — authenticate and authorize requests made with your API keys.
  • Authentication — verify your identity and manage session security.
  • Sharing & review — generate share links and grant guest reviewers access to the specific content you share.
  • Transactional email — send account, invitation, and notification emails.
  • Analytics — understand usage and improve the service.

4. Third-Party Services

We use the following third-party services to operate Letter:

  • Supabase — database hosting, authentication, and file storage. Your data is stored in Supabase-managed PostgreSQL databases.
  • Vercel — application hosting and deployment, including hosting for the blogs and sites you deploy and connecting their domains.
  • Anthropic (Claude API) — AI processing of text for suggestions, analysis, and visualization generation, and of repository content for the code changes you ask for. Content sent to Claude is subject to Anthropic's Privacy Policy.
  • OpenAI / Bloom — AI image generation, depending on our configuration.
  • OpenAI — AI processing of repository content for code changes, when the change is made with an OpenAI model.
  • Concentrate — the gateway through which we send requests to Anthropic and OpenAI models, including requests that carry repository content.
  • Google — Sign in with Google (OAuth) for authentication, Google Docs as an optional publishing destination, and Gmail (the Gmail API over OAuth, send and read-only scopes) for mailboxes you choose to connect.
  • Notion, Contentful, Webflow, Sanity, and Framer — optional content integrations for importing and publishing content. Connected via OAuth or API key, with encrypted credential storage.
  • GitHub and GitLab — optional connections to your code repositories, through the Letter GitHub App, GitLab OAuth, or an access token you provide.
  • Maritime — the temporary, isolated build environments in which code changes are made and previewed.
  • Lemma — a record of each code change's steps: the instructions, the agent's messages, the tools it used and their results (which can include excerpts of files it read), and token counts. We use these records to find and fix failures in how changes are built.
  • PostHog — product-usage analytics.
  • RB2B — identification of business visitors to our website, for sales outreach. See RB2B's Privacy Policy.
  • Resend — delivery of transactional email.

5. Data Security

  • Integration credentials (OAuth tokens and API keys) are encrypted at rest using AES-256-GCM with unique initialization vectors per credential.
  • Letter API keys are stored only as salted hashes — never in plaintext.
  • Database access is protected by row-level security policies, ensuring users can only access data belonging to their organizations.
  • All data is transmitted over HTTPS.
  • Passwords are managed by Supabase Auth and are never stored in plaintext.
  • API endpoints are rate-limited to prevent abuse.

6. Data Sharing

We do not sell your personal information. We share data only in the following circumstances:

  • AI processing — content is sent to our AI providers when you use enhancement or generation features.
  • Publishing & sync — content is sent to a connected platform when you import from or publish to it.
  • Code changes — repository content is processed in a temporary build environment and sent to our AI providers when you ask for a change to a connected repository.
  • Deployments & domains — deployment content is hosted by Vercel, and domain-registration details are shared with the registrar when you buy a domain.
  • Shared links & guest reviewers — when you enable a share link or invite a guest reviewer, the shared content is accessible to anyone with the link (and, for reviewers, the associated PIN).
  • Service providers — analytics and email vendors process data on our behalf to operate the service.
  • Legal requirements — we may disclose data if required by law, legal process, or government request.

7. Data Retention

Your data is retained for as long as your account is active. Version history is maintained indefinitely to support content versioning. If you delete your account, we will remove your personal data from our systems, though some data may persist in backups for a limited period.

Records of code changes (instructions, summaries, changed-file lists and agent logs) are kept while your account is active. Disconnecting a repository stops our access but does not delete these records; to delete them without deleting your account, contact us at the email address below.

8. Your Rights

You may:

  • Access, update, or delete your account information through your profile settings.
  • Disconnect any content integration at any time through your organization settings.
  • Disconnect GitHub or GitLab, or uninstall the Letter GitHub App, at any time.
  • Revoke an API key at any time.
  • Disable a share link or revoke a guest reviewer's access at any time.
  • Request deletion of your data by contacting us at the email address below.

9. Changes to This Policy

We may update this Privacy Policy from time to time. We will notify you of significant changes by posting the updated policy on this page with a revised date. Your continued use of Letter after changes are posted constitutes acceptance of the updated policy.

10. Contact

If you have questions about this Privacy Policy or your data, contact us at privacy@letterbrace.com.