Privacy Policy

1. Scope

This policy covers the Records Labs platform: the app, APIs, website widgets, and connected-service integrations. The browser extension has its own, more specific policy: Extension Privacy. When your organization uses Records Labs, your organization is the controller of the content it brings and of its website visitors’ data; Records Labs processes that data to provide the Service.

2. What we collect

Account and workspace data

• Name, email, and authentication data for signing in.
• Workspace configuration: members, roles, agents, knowledge rules, integrations.

Customer content

• Documents, pages, transcripts, and records your organization uploads, captures, or connects (for example from Google Drive, Gmail, Slack, HubSpot, Epicor). We process it (extraction, chunking, embedding, transcription) to answer questions with citations.
• Questions asked and answers generated, with usage and quality telemetry.

Website widget visitors

• Messages a visitor sends to a widget, limited page context (page URL/title where the widget is embedded), and technical data (IP-derived rate-limit keys, origin).
• Identity only if the embedding site provides it: self-reported details (such as a name or email typed into a form) or a cryptographically signed identity from the site’s own login system, optionally including traits (such as a plan or tier) used to scope which knowledge the widget may use.
• Widgets support a consent-required tracking mode; the embedding organization is responsible for its own visitor notice and consent.

Billing

• Payments run through Stripe. We store plan, credit, and invoice records; we never store full card numbers — Stripe holds payment credentials.

3. How we use data

• To provide, secure, and improve the Service (including abuse and fraud prevention).
• To generate answers: relevant content is sent to large language models acting as our processors. By default, answers are written only by model providers that do not train on customer content. An organization administrator can change this in the organization’s settings.
• To read and index your files: we use the document-reading and search providers listed in section 4. The organization setting above does not apply to them. They handle content under their own terms, and some of those terms allow the provider to use content to improve its own services.
• To meter usage, bill credits, and send operational notices.
• To comply with law and enforce our Terms.

4. Subprocessors

We use a small set of infrastructure providers to run the Service:

• Supabase — database, authentication, and file storage.
• Railway — backend compute.
• Vercel — web application hosting.
• Stripe — payments and billing.
• OpenRouter and the model providers it routes to (for example Anthropic, OpenAI, Google, xAI) — AI answer generation, and summaries, tags, and image descriptions while documents are processed.
• Google (Gemini API) — reading figures, charts, page images, standalone images, and audio and video recordings in your files.
• Mistral AI — text recognition for PDF pages.
• Reducto — detailed reading of complex PDF pages, such as dense tables and drawings.
• RunPod and Modal — computing capacity that runs our own text-recognition model on PDF pages. They receive page images through short-lived links, and we delete those images after processing.
• Voyage AI — search indexing and result ranking for your content.
• turbopuffer — search index storage (numeric representations and record IDs, not the text of your documents).
• OpenAI — transcription of audio and video recordings.
• Exa — web search, when an answer draws on public web results.
• Resend — sending email, and receiving email sent to a workspace’s inbound addresses.
• Google APIs — only for workspaces that connect Google Drive/Gmail sources.
• Twilio — only for workspaces that enable SMS/voice channels.

These providers may process data in the United States and in other countries where they operate. Each processes data only as needed to provide its function. A current list is available on request at privacy@recordslabs.ai.

5. Retention and deletion

• Customer content follows your workspace’s retention and trash controls; deleted content is purged on the schedule shown in the app.
• Query and audit logs are retained for operations, security, and billing accuracy, then deleted or de-identified.
• On workspace termination, we delete or de-identify customer content within a commercially reasonable period, except where law requires retention.

6. Security

Tenant isolation, role- and rule-based access controls, encryption in transit, secret redaction on sensitive fields, and audit logging. Widget knowledge exposure is fail-closed: public widgets see only knowledge explicitly marked public unless an administrator configures a stronger trust anchor (origin lock, signed page pass, or signed-in visitor identity).

7. Your rights

Depending on your jurisdiction, you may have rights to access, correct, export, or delete personal data. Workspace members should contact their administrator first (most actions are self-serve); anyone can reach us at privacy@recordslabs.ai. Widget visitors should contact the organization whose site embeds the widget — that organization controls their data — and we will support its requests.

8. Children

The Service is a workplace tool and is not directed to children under 13.

9. Changes

We may update this policy; material changes will be announced in the app or by email and reflected here with a new date.

10. Contact

Records Labs — privacy@recordslabs.ai