Privacy policy.

This policy covers the Amoeba desktop app, the service it syncs to (we call it the Brain) and the website at useamoeba.com. Amoeba, Inc., a Delaware corporation ("Amoeba", "we" and "us"), is the controller of the personal data described here. For general, legal and privacy requests, write to ken@useamoeba.com.

When you sign in

You can sign in with an email one-time code, Google or GitHub. We store the account identifiers and profile information returned by your selected sign-in method, including your email address and any display name and avatar. We use them to identify you, link sign-in methods you verify, and control workspace access. Connecting GitHub for repository access is separate from email or Google sign-in. When you connect GitHub, we store your GitHub access token, encrypted, to give you repository access. We do not store your local Git credentials or SSH keys.

Inside a workspace

A workspace belongs to a team. Session metadata (who started a session, when, its task, the files it touched, its status) and session transcripts and context that are shared inside the workspace are synced to the Brain so your teammates can see them. Transcripts can include snippets of code your agent read or wrote.

• Transcript chunks and stored GitHub tokens are encrypted by Amoeba before they are written to the database. Other workspace records are protected by our database provider’s storage encryption. Amoeba holds the encryption keys, so this is not end-to-end encryption.
• Amoeba staff do not access workspace content except to give support you ask for, to investigate security or abuse, or where the law requires.
• Amoeba does not train or fine-tune models on your prompts, code or transcripts. Amoeba does use models to process your workspace content to produce task summaries, overlap checks and memories. Third-party providers process data under their own terms.
• The sub-processors listed below process data on our behalf to run the service.

Local credentials and repositories

Claude Code and Codex run on your machine, and their configured credentials are managed by the local CLI. Repository content remains in local checkouts and moves through your configured Git remote under your Git credentials. The Brain stores snapshot commit and ref metadata, not a repository copy.

Prompt processing

When you use your own provider account, your selected coding provider processes prompts and necessary context under that account’s data controls. With Amoeba-managed model usage, prompts, selected workspace context and tool results pass through the Brain to the selected model provider, directly or through Straitly when that route is enabled. Provider and supplier processing is subject to their applicable data terms. Amoeba also processes prompt-derived coordination context through the Brain. Anthropic can process that content for overlap checks and task summaries. Built-in pattern-based secret redaction is applied before transcript chunks are stored, but you should still avoid putting secrets in prompts.

Managed usage and payments

We store usage, model and supplier identifiers, cost and allowance records, payment status, Stripe customer and payment-method identifiers, invoices and consent records to authorize requests, collect payment and reconcile billing. Stripe collects and stores payment details through its hosted setup and billing pages; Amoeba does not receive your full card number. Optional Link use is governed by Stripe’s terms and privacy policy.

Error reporting

Automatic error reporting is on by default to help us diagnose startup failures, crashes and missing interface resources. Reports are linked to your signed-in Amoeba account and sent to the Brain for Amoeba support. You can turn reporting off with the “Share error reports with Amoeba” switch during setup.

• Reports contain structured diagnostics such as app version, operating system and architecture, failure type, diagnostic codes and locations in Amoeba’s own source files.
• Reports exclude prompts, your code, terminal output, credentials, arbitrary error messages, screenshots, browsing URLs and memory dumps. This setting does not enable the underlying editor’s telemetry.
• Turning reporting off stops automatic collection and uploads and deletes unsent reports on this device. Your preference syncs to your account when connected and signed in.
• Unsent reports are limited to 50 events for up to seven days on your device. Reports received by the Brain expire after 30 days and are deleted when you delete your account.

Early access signups

If you give us your email address for early access, it is stored in our database, and we send a confirmation with Resend. We use it to get in touch about Amoeba and for nothing else.

Analytics

The site uses first-party, cookieless analytics. For each page view we record the page path, the referrer, the browser family and the operating system, plus a hash of your IP address and user agent. The hash rotates daily and cannot be reversed, so it cannot be used to identify you or to follow you from one day to the next. There are no cookies and no third-party trackers. If your browser sends the Global Privacy Control signal, we honour it.

Discord

Our community runs on Discord. Anything you do there is covered by Discord's own privacy policy, not this one.

We do not sell data and we do not run ads. These providers process data on our behalf, only to run the service:

We may also disclose data if the law requires it.

• Providing the service. We use account, workspace and billing data to give you Amoeba and to bill managed usage. This is the contract between you and us.
• Security and diagnostics. We use error reports, abuse checks and website analytics for our legitimate interests in keeping Amoeba secure and working.
• Email and signups. We send optional marketing email and keep early access signups only with your consent. You can withdraw it at any time by writing to ken@useamoeba.com.

• Transcript chunks follow your organization’s configured transcript-retention period. The default is 90 days, and an organization can change it. Other coordination records do not share that automatic expiry.
• Deleting your account anonymises identity fields and redacts associated transcript chunks, subject to account requirements. It does not erase every collaboration record.
• Deleting a workspace removes workspace-scoped Brain data. Local Amoeba state, Git objects, local provider history and provider-hosted history are managed separately.
• Early access signups are kept until you ask us to remove them.
• The analytics hash rotates every day, so past page views cannot be tied back to you.

Amoeba hosts and processes data in the United States. Where we transfer data from the EU or UK, we rely on standard contractual clauses where required.

You can ask what we hold about you, ask us to correct it, or ask us to delete it. If you are in the EU, the UK or California, you also have the right to access, correct, delete and port your data. To ask, write to ken@useamoeba.com. We respond within 30 days, a person answers, and we will confirm when it is done.

Amoeba is not directed at children under 16, and we do not knowingly collect data from them. If you think a child has given us data, tell us and we will delete it.

We post changes on this page with a new effective date. For material changes we also announce them in the changelog and on Discord.