Privacy Policy

Effective date: 2026-06-01 · Last updated: 2026-06-01

Anrie is operated by Unreal Labs, Inc. ("Unreal Labs", "we", "us", or "our"). This Privacy Policy explains what information we collect when you use Anrie (the "Service", available at anrie.ai), how we use and protect it, and the choices and rights you have.

Anrie is a business product designed to be used by companies and their team members. If you use Anrie on behalf of an organization, that organization (your "Workspace") is the controller of the data you submit through the Service, and we act as a processor on its behalf.

If you have any questions about this policy, contact us at [email protected].

1. Who We Are

2. What Anrie Does

Anrie is an AI "coworker" that operates primarily inside Slack. It connects to your Slack workspace and to your advertising, commerce, and productivity tools (for example Meta Ads, Google Ads, Shopify, HubSpot, Notion, and Google Sheets) to help your team generate, review, and manage ad creatives and related work.

To do this, Anrie needs to authenticate to those platforms on your behalf and process the data you direct it to work with. The sections below describe exactly what we collect, what we store, and what we only handle transiently.

3. Information We Collect

3.1 Slack authorization and identity data

When you install Anrie into a Slack workspace via Slack OAuth, we collect and store:

• Slack OAuth bot token (the xoxb- bot access token issued by Slack). This is used to make API calls to Slack on your workspace's behalf.
• Slack workspace identifiers: Slack team ID, enterprise ID (if applicable), team/workspace name, bot ID, and bot user ID.
• Anrie user identity: your email address, your Slack user ID, and your workspace membership.

This data is stored in our application database (anrie_workspaces and anrie_users) and is retained for as long as your workspace remains installed (see Section 8, Data Retention).

3.2 Connected ad-account and platform credentials (via Pipedream)

When you connect a third-party platform (Meta Ads, Google Ads, Shopify, HubSpot, Notion, Google Sheets, and similar), the OAuth authorization is brokered by Pipedream. Importantly:

• The actual access tokens for these platforms are held by Pipedream, not stored by Anrie.
• Anrie stores only references and metadata: a Pipedream account identifier, the app slug (e.g. facebook_graph_api), the authorized scopes, and non-sensitive account metadata (such as the account name and connection health status).
• We also store a mapping between your Workspace and your Pipedream connection (an external user ID of the form anrie_workspace:<workspaceId>, plus the Pipedream project and environment).

3.3 Settings, configuration, and integration data

• Workspace settings and resource selections: which ad accounts, campaigns, spreadsheets, or other resources you have selected for Anrie to use (resource type, external ID, display name, and associated metadata).
• Integration catalog: read-only reference data describing available integrations (app slugs, display names, scopes, capabilities). This is reference data, not personal data.
• Integration event payloads: webhook events received from Pipedream (for example account-update or sync events). These are stored for audit and debugging. They contain Pipedream account/sync data and do not contain raw Slack message content or files.

3.4 Generated outputs and creative work

Content that Anrie generates or works with at your direction — such as ad copy, generated creatives, briefs, reports, and related artifacts — together with the inputs you provide (prompts, uploaded assets, reference materials). The location and retention of these outputs depend on where they are produced; some are stored in our object storage and analytics systems described in Section 6.

3.5 Authentication session data

When you sign in, we issue a session token (a signed/encrypted NextAuth JWT) stored in an httpOnly browser cookie. The token contains your user ID, workspace ID, Slack user ID, Slack team ID, and email. It does not contain your Slack bot token.

3.6 Service logs and operational data

• Error and performance monitoring: errors, stack traces, and request/runtime context captured by Sentry to operate and debug the Service. This context may incidentally include identifiers (such as workspace or user IDs) present in the error. Sentry session replay is sampled. Sentry is configured with default PII protections (sendDefaultPii disabled; session replay masks text and input content) and retains events for within 90 days.
• Infrastructure logs and metrics: operational logs, application metrics, and analytics records used to run, monitor, and secure the Service.

3.7 Cookies

Anrie uses only strictly necessary cookies required to authenticate users and operate the Service. The primary cookie is an httpOnly NextAuth session cookie containing a signed/encrypted token (see Section 3.5). We do not use cookies for advertising, third-party analytics, or cross-site tracking.

4. How Slack Data Is Handled: Stored vs. Transient

We treat your Slack data with a strict data-minimization approach. This section is intentionally explicit.

4.1 Slack data we store

• The Slack OAuth bot token (so Anrie can act in your workspace).
• Slack workspace and user identity (team ID, enterprise ID, bot ID/user ID, and per-user email and Slack user ID).
• Channel provisioning and membership state required to operate the integration — for example which channels Anrie is a member of and pagination cursors used to sync. This is configuration/state, not message content.

4.2 Slack message content and conversation history

When Anrie operates in a Slack channel, it stores messages, threads, and metadata from that channel so the agent can maintain conversation context across sessions. This includes message text, author and channel identifiers, timestamps, and thread relationships.

Anrie only stores Slack content from channels where the Anrie bot has been added. Anrie does not access or store content from channels where it has not been invited, even though the Slack OAuth scopes it holds (such as channels:history, users:read.email, and files:read) would technically permit broader access.

Slack content is stored on our EU-hosted infrastructure (Google Cloud, europe-west1). Tokens, signing secrets, and authorization data on raw events are stripped before storage. An archive is retained in our BigQuery analytics warehouse in the same region for reconciliation purposes.

Retention and deletion of stored Slack content: Stored Slack content is retained while Anrie is installed in your workspace. Within 30 days of app uninstall, account closure, or a verified deletion request, this content is deleted from production systems (see Sections 8 and 9). You can request deletion at any time — including while your installation remains active — by emailing [email protected].

4.3 Connected-platform data

Data fetched from your connected platforms (ad accounts, campaigns, performance metrics, products, CRM records, sheets, documents) is brokered through Pipedream and processed by Anrie's agent to do the work you request. The underlying OAuth tokens for those platforms are held by Pipedream, not by Anrie (see Section 3.2).

5. How We Use Your Information

We use the information described above to:

• Provide, operate, and maintain the Service, including authenticating to Slack and your connected platforms on your behalf;
• Power Anrie's AI agent so it can read the context you direct it to and generate ad creatives, copy, reports, and related outputs;
• Sync and display your connected accounts, settings, and selected resources;
• Monitor, debug, secure, and improve the Service (including error tracking and performance monitoring);
• Communicate with you about the Service and respond to support and privacy requests;
• Comply with legal obligations and enforce our terms.

We do not sell your personal information.

6. AI and Large Language Models (LLMs)

Anrie's agent reasoning is powered by large language models. When Anrie processes your data to plan and generate work, the relevant content (such as prompts, Slack context you direct it to, connected-platform data, and creative inputs) is sent to LLM providers to produce a response.

• Primary reasoning model: Anthropic's Claude, served through Google Vertex AI within our Google Cloud project. For the core Anrie agent, customer data sent for reasoning flows to Google Cloud / Vertex AI rather than directly to Anthropic's public API. Vertex AI is configured for multi-region (global) routing, so prompts and the Slack and connected-platform context Anrie reasons over may be processed outside the EU (see Section 11).
• Creative-generation models: When you request media or creative generation, additional AI providers may process your creative inputs and produce outputs. Depending on the requested task these can include OpenAI, Google Gemini/Vertex AI, xAI (Grok), Fal.ai, ElevenLabs, Hedra, and HeyGen.

A current list of AI providers and other subprocessors is maintained on our Subprocessors page.

No training on customer data

We do not use your content to train, fine-tune, or build Anrie or any third-party foundation models. Your content is used solely to provide the Service to you and your Workspace. Where we use third-party AI providers, we rely on contractual and technical controls intended to prevent training on your content. Specific retention practices vary by provider — providers may, for example, retain inputs for a short period for abuse monitoring or safety review as described in their own terms. Current provider practices are summarized on our Subprocessors page and updated as provider terms change.

7. Where Your Data Is Hosted and Subprocessors

7.1 Hosting

Anrie runs on Google Cloud Platform (GCP). Our primary data stores are located in the europe-west1 region (EU). This includes:

• Application database — Google Cloud SQL for PostgreSQL (europe-west1).
• Analytics warehouse — Google BigQuery (europe-west1).
• Object storage — Google Cloud Storage buckets located in the EU / europe-west1 region.

Public web and Slack traffic is fronted by Cloudflare, which provides DNS, TLS termination, CDN/edge, and DDoS/WAF protection for anrie.ai.

Some object-storage buckets use the multi-region EU location, which may span more than one EU country. We do not currently offer single-country data residency.

7.2 Subprocessors

We use third-party subprocessors to provide the Service, including cloud hosting, AI/LLM providers, the Slack integration, the Pipedream integration broker, error monitoring, and CDN/edge services. We maintain a current list of subprocessors, including their purpose and location, on our Subprocessors page.

8. Data Retention

• Account, workspace, and connection data (Slack identity and bot token, user records, Pipedream account references, settings, integration catalog and resources) is retained for as long as your Workspace's installation remains active, and thereafter until deleted per Section 9.
• Slack conversation history and the communications ledger are retained by the UCS backend while your installation is active, and are deleted within within 30 days after you uninstall the Anrie app, close your account, or make a valid deletion request (see Sections 4.2 and 9).
• Temporary uploads and caches age out automatically. For example, temporary upload buckets auto-delete on a short cycle (e.g. ~30 days), shared workflow upload storage auto-deletes on a longer cycle (e.g. ~90 days), and thumbnail/agent caches expire on their own lifecycle. Specific lifecycle windows vary by storage class and bucket.
• Error/monitoring data is retained per our Sentry configuration (within 90 days).

Deletion timeline

When you uninstall the Anrie app, close your account, or make a valid deletion request (Section 9), we will delete the relevant personal data — including Slack conversation history and the communications ledger — from our production systems within within 30 days. Our database backups (point-in-time recovery and automated backups) are retained on a rolling basis and age out within within 14 days thereafter, after which deleted data is no longer present in backups. During that interval, deleted data may persist only in encrypted backups and is not used for active processing.

Backup retention reflects the current Cloud SQL configuration: 7-day point-in-time recovery plus 14 retained automated daily backups.

9. How to Request Access or Deletion

To request access to, correction of, or deletion of your personal data, email [email protected] from the email address associated with your Anrie account, or have a Workspace administrator contact us on your behalf.

Please include enough information for us to identify the relevant records, such as:

• Your name and email,
• Your Slack workspace / team identifier (or workspace name), and
• The nature of your request (access, correction, deletion, etc.).

What happens after a deletion request:

1. We verify the request and your relationship to the Workspace.
2. We delete the relevant personal data from our production systems (application database, Slack history stored by UCS, the communications ledger, and associated object storage) within within 30 days.
3. Residual copies in encrypted backups age out within within 14 days of deletion, as described in Section 8.
4. We instruct relevant subprocessors to delete the corresponding data where applicable, and disconnect connected platform accounts brokered through Pipedream.

Because Anrie is typically used on behalf of an organization, some requests may need to be directed to, or confirmed by, your Workspace administrator (the data controller). We will respond to verified requests within the timeframes required by applicable law.

10. Your Rights

Depending on where you live, you may have rights regarding your personal data.

10.1 EEA / UK (GDPR)

If you are in the European Economic Area or the United Kingdom, you may have the right to:

• Access the personal data we hold about you;
• Rectify inaccurate or incomplete data;
• Erase your data ("right to be forgotten");
• Restrict or object to certain processing;
• Data portability;
• Withdraw consent where processing is based on consent;
• Lodge a complaint with your local supervisory authority.

The legal bases on which we process personal data include performance of a contract, our legitimate interests in operating and securing the Service, your consent (where applicable), and compliance with legal obligations.

10.2 California (CCPA/CPRA) and other US state laws

If you are a California resident, you may have the right to know what personal information we collect, to request deletion, to correct inaccurate information, and to not be discriminated against for exercising your rights. We do not sell your personal information and do not "share" it for cross-context behavioral advertising as those terms are defined under California law. When you direct Anrie to send data to an advertising platform you have connected (such as Meta Ads or Google Ads), Anrie acts as your service provider on your instruction; that transfer is not a "sale" or "share" by Anrie.

To exercise any of these rights, contact [email protected] (see Section 9). You may use an authorized agent where permitted by law.

11. International Data Transfers

Our primary infrastructure is located in the EU (europe-west1). However, some of our subprocessors — including certain AI/LLM providers and operational services — may process data in other countries, including the United States. In particular, LLM inference for the Anrie agent uses Google Vertex AI in a multi-region (global) configuration and may be processed outside the EU. The Slack platform, Pipedream, and your connected platforms may also process data outside the EU depending on their own configurations.

Where personal data is transferred outside the EEA/UK, we rely on appropriate safeguards, such as the European Commission's Standard Contractual Clauses, where required. For such transfers we rely on the European Commission's Standard Contractual Clauses and, for UK transfers, the UK International Data Transfer Addendum, together with any adequacy decision or data-transfer framework a given subprocessor maintains.

12. Security

We take security seriously and apply technical and organizational measures to protect your data, including:

• Encryption in transit: TLS is enforced for traffic to the Service. Cloudflare is configured for strict SSL, TLS 1.3, and always-use-HTTPS. Database connections require TLS (sslmode=require), and internal service-to-service traffic runs over TLS.
• Encryption at rest: Data stored in Cloud SQL (PostgreSQL), Cloud Storage, and BigQuery is encrypted at rest using Google-managed encryption keys. Backups are encrypted. (We currently use Google-managed keys rather than customer-managed keys. Customer-managed encryption keys (CMEK) are not currently offered.)
• Network isolation: Production runs in a private GCP VPC with private subnets; database access is over private IP; the Kubernetes cluster has no public node IPs; and public ingress is restricted to Cloudflare-fronted endpoints.
• Secrets management: Credentials are managed in a dedicated secrets store (1Password as the source of truth, synced to GCP Secret Manager) with IAM-scoped access; the Slack signing secret and OAuth credentials are stored as managed secrets. Inbound Slack webhooks are verified using a signing secret.
• Access controls: Internal/administrative surfaces are gated behind Cloudflare Access with an allowlist for team members; access to systems containing personal data is restricted to authorized personnel.
• Token handling: Slack bot tokens are stored server-side and are never exposed to client-side code or browser sessions; connected-platform tokens are held by Pipedream, not by Anrie.
• Monitoring: We use error tracking, logging, and alerting to detect and respond to operational and security issues.

We are an early-stage company and do not yet hold third-party security certifications such as SOC 2 or ISO 27001. Single sign-on (SSO/SAML) for enterprise accounts is on our roadmap. Customers with formal compliance requirements may request a security overview at [email protected], and we will work with you to address specific questions, controls, or due diligence questionnaires.

No method of transmission or storage is completely secure, and we cannot guarantee absolute security. If we become aware of a security incident affecting your personal data, we will notify affected parties as required by applicable law. To report a security concern, contact [email protected].

13. Children's Privacy

Anrie is a business product intended for use by organizations and their team members. It is not directed to individuals under 16 years of age, and we do not knowingly collect personal data from anyone under 16. If you believe a minor has provided us with personal data, contact [email protected] and we will take steps to delete it.

14. Changes to This Policy

We may update this Privacy Policy from time to time. When we make material changes, we will update the "Last updated" date at the top of this page and, where appropriate, provide additional notice. Your continued use of the Service after an update constitutes acceptance of the revised policy.

15. Contact Us

For all privacy, security, and legal inquiries, contact [email protected].

Unreal Labs, Inc. — operator of Anrie (anrie.ai) Headquarters: 2261 Market St, San Francisco, CA 94114, USA · Registered: 131 Continental Dr, Suite 305, Newark, DE 19713, USA

London Unreal Labs Ltd — UK affiliate (company number 16697867) 3rd Floor 1 Ashley Road, Altrincham, Cheshire, WA14 2DT, United Kingdom