Afora Privacy Policy

Effective date: July 10, 2026 · Last updated: August 29, 2026

Afora LLC ("Afora," "we," "us") provides a hosted personal AI agent: your own agent, running around the clock on its own dedicated instance, that you reach from a browser console and the Afora Mac app, and that runs on the AI subscription you already have (Anthropic Claude or OpenAI ChatGPT). This policy explains what data we handle, how we use it, and the choices you have. It is written to be read; if anything is unclear, email privacy@aforademo.com and we will answer plainly.

This policy is incorporated into our Terms of Service. Our Security page describes how we protect data; our Subprocessor list names every vendor that touches customer data.

1. What Afora is, in data terms

When you subscribe, we provision an agent instance for you: an isolated environment on infrastructure we operate, with its own workspace. Your conversations with your agent, the files in its workspace, its memory, and its activity all live on that instance. The web app at aforademo.com is the front door — signup, billing, and the handoff into your console.

You are our customer directly. If you use Afora inside a company, your use may also be governed by your company's own policies; data your agent touches in accounts you connect (for example, your email) remains governed by your relationship with those account providers.

2. Information we collect

We do not collect: passwords for your AI provider or connected accounts (sign-in happens with the provider and we receive only a token), biometric data, precise geolocation, or data about children. Afora is not directed at anyone under 18.

3. How we use information

We use data only to provide and secure the service:

• Run your agent: relay your instructions to it, run the work you ask for, and show you the results in the console and Mac app.
• Operate your instance: provision it, keep it running, restore it after failures.
• Bill your subscription and communicate with you about the service.
• Operate, secure, debug, and improve the service (rate limiting, fraud and abuse detection, error diagnosis).

We do not use your data for advertising of any kind, and we do not sell personal information — to anyone, for anything.

4. Artificial intelligence — your own subscription

This is the most important section of this policy, because it is different from most AI products: your agent thinks with the AI subscription you connect — your Anthropic Claude account or your OpenAI ChatGPT account (or an Anthropic API key you provide).

• Your agreement governs the AI processing. When your agent sends a prompt, the content goes to the provider you connected, under your agreement with that provider and your settings in that account — not under a contract between Afora and the provider. Check your provider's data controls (for example, model-training settings in your ChatGPT or Claude account) — those settings, not this policy, decide what the provider does with that content.
• What we store. We store the sign-in token for the provider you connect, encrypted, scoped to your instance, so your agent can work. Disconnecting the provider in the console retires the token.
• Afora does not train models on your data. We do not use your conversations, workspace, or connected-account data to train AI or machine-learning models, and we do not permit our own subprocessors to do so.
• Isolation. Your agent works only with your instance's data. One customer's data is never used in another customer's sessions.

5. Connected accounts — Google and Microsoft commitments

Connecting any outside account is optional and uses OAuth: you grant access from the provider's own consent screen, and you can revoke it at any time from the provider's security settings or from within Afora.

Google. Afora's use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements. In particular:

• We use Google user data only to provide the user-facing features you invoke through your agent — never for advertising, and never to determine creditworthiness or for lending purposes.
• We do not transfer Google user data to third parties except to provide those features (see our subprocessor list and Section 4 on your connected AI provider), for security purposes, to comply with applicable law, or as part of a merger or acquisition with notice to you.
• We do not use Google user data to develop, improve, or train generalized (non-personalized) AI or machine-learning models.
• No human at Afora reads your email or file content, except: (a) with your explicit permission (for example, a support request where you share a specific item); (b) as necessary for security purposes such as investigating abuse; or (c) as required by law.

Microsoft. We apply the same commitments to Microsoft 365 / Outlook data accessed through Microsoft Graph: minimum necessary permissions, no advertising or marketing use, no use outside the permissions you granted, and retention and deletion per Section 8.

Other connections (for example GitHub, Slack, Linear, Telegram) follow the same pattern: your agent accesses them only with the token you granted, only to do the work you direct, and you can disconnect at any time.

6. Your agent acts for you

Afora is an agent, not just a chat window: when you direct it to (and within the connections you have granted), it can send messages, edit files, and act in your connected accounts. Content it produces and actions it takes are attributed to you. The console shows its activity; review what matters. Content your agent encounters in the outside world (web pages, inbound messages) is treated as data, and manipulation attempts (prompt injection) are a threat we design against — see the Security page.

7. When we share information

We share personal information only with:

• Service providers (subprocessors) that host and power Afora — currently Vercel (web app hosting), Hetzner (agent instance hosting, EU), Supabase (database), Clerk (authentication), Stripe (billing), and the operational vendors on our Subprocessor list. Each is bound by contract to use data only to provide its service to us. We give customers at least 30 days' notice before adding or replacing a subprocessor that handles customer data.
• The AI provider you connect (Anthropic or OpenAI) — under your own agreement with them (Section 4).
• The services you connect — only as your agent acts in them at your direction (Section 5).
• Professional advisors (lawyers, accountants) under confidentiality obligations.
• Authorities, when required by law, subpoena, or to protect rights and safety — and where lawful, we will notify the affected customer first.
• A successor entity in a merger, acquisition, or asset sale, with notice to you.

We never sell or rent personal information, and we never share it with data brokers or advertising platforms.

8. Retention and deletion

• Workspace and conversation data lives on your instance while your subscription is active.
• On cancellation: you may export your data for 30 days after the subscription ends; we then delete your instance's data from production systems within 60 days, except minimal records required for legal, tax, or audit purposes. Backups age out on a fixed schedule (see Security page).
• Connected-account tokens are retired when you disconnect the account, and deleted with the instance.
• AI provider retention for content your agent sent to your connected provider is governed by your agreement and settings with that provider (Section 4).

9. Your rights and choices

• Disconnect at any time: your AI provider and every connected account can be disconnected in the console or revoked at the provider.
• Access, correction, export, deletion: email privacy@aforademo.com and we will honor reasonable requests to access, correct, export, or delete your personal information, regardless of whether a specific privacy statute requires it.
• Marketing: we send only service and account emails; there is no ad targeting to opt out of.

10. Security

Traffic is encrypted in transit (TLS 1.2+). Each customer's agent runs in its own isolated instance with its own operating-system identity and data directory — one customer's agent cannot read another's data. Credentials and tokens are encrypted at the application layer where our web app stores them, and held on your instance under its own isolation elsewhere. Full details, including how to report a vulnerability, are on our Security page. No system is perfectly secure; if a breach affects your data, we will notify affected customers without undue delay, and within 48 hours of confirmation for personal-data breaches.

11. Changes to this policy

We will post changes here with an updated date. If a change materially expands how we use personal information, we will notify affected customers in advance and, where the change involves using existing data for a new purpose, obtain consent before applying it. We will never quietly repurpose your data through a policy edit.

12. Contact

Afora LLC · privacy@aforademo.com