Privacy policy

Information we collect and why

• Identity and access: name, email, profile image, account IDs and usernames from you or Google/GitHub sign-in; a password hash if you use a password; invitations, memberships and device authorization records. We use these to create accounts, authenticate you, control access and prevent abuse.
• Your work: repositories you connect, source files, workspace disks, imported conversations, prompts, chat history, attachments, agent output, commands and collaboration activity. We process these to run your workspaces and the features you request.
• Connections and secrets: provider access/refresh tokens, API keys, official CLI authentication material and environment variables you supply. Stored credential material is encrypted; authorized backend and runtime processes need access to use your connections. Only connect accounts and repositories you are authorized to use.
• Billing: Stripe customer and subscription IDs, plan, payment status and billing dates, and for Teams the company's name and what it does. Stripe collects payment and billing details directly; CoDev does not receive full card numbers. We use billing records to provide paid access, handle refunds, accounting and disputes, and company details to understand which kinds of teams use CoDev.
• Support and waitlist: contact information, feedback, access requests (including the name and email Google or GitHub shares when you request access with them) and information you send us, used to respond, invite you and operate the service. Email service providers process delivery and correspondence.
• Technical and usage information: requests, IP addresses visible to our infrastructure, security logs, errors, browser/device information, compute usage and setup milestones (such as creating a workspace or running a first agent), which drive your setup checklist and product-wide totals. With optional analytics enabled, we also record page visits, referrer origin and an IP-derived hash, which may be associated with your signed-in account, and a short campaign source such as Discord, Reddit or X when you sign up. We use this to understand usage and improve the service.

AI providers and shared workspaces

When you run an agent, relevant prompts, files, conversation history and tool results are sent to the provider selected for that task, such as OpenAI, Anthropic, Cursor, Microsoft Azure AI Foundry or Amazon Bedrock, including through accounts you connect. Their handling, retention and any training use depend on that provider, product and your account settings. Review those terms before sending confidential or personal information. We do not promise that third-party services retain no data.

Members with access to a shared workspace or chat can see content available there, including agent output and files. Agents can execute commands and contact external services. Do not put secrets or information you cannot share in shared content. Removing a member cannot recall copies others already downloaded.

Administrator and staff access

CoDev administrators manage accounts and plans from an internal console. It shows account details (account ID, name, email and whether it is verified, username, sign-in methods, plan and join date; with optional analytics, visit activity), waitlist requests, and product-wide totals such as how many workspaces exist, which AI providers are connected and used, how many credits were granted and spent across all accounts, how many operations failed, and how many members reached setup milestones, usage trends and anonymous activity rankings. So that we can control costs, stop abuse and answer billing questions, it also shows each account's usage: how many workspaces it owns and their state (running, stopped or failed), whether they are shared, how many invites it sent, how many hours its virtual machines ran, and its credit balance, spending and recent credit history; and whether it has connected its own Claude, ChatGPT or Cursor subscription (which provider and when, never the sign-in itself). Workspaces appear only by an internal ID, never by name. It also lists reported bugs, with the error message redacted as described below and without who reported them. It has no feature to open a workspace or view its name, repository, files, chats, terminals, Git history, members, environment variables or connected credentials, and administrator status does not make anyone a member of your workspace. Administrators can grant credits to an account and take back unspent credits that administrators granted. Each change an administrator makes to a plan, subscription, organization role, administrator access, credits or beta access is recorded with who made it and the old and new values, and administrators can review that record in the console.

Saved credentials and environment variables are encrypted with keys held in a managed key service and are never displayed again, to you or anyone else, after you save them. CoDev decrypts them only to connect your accounts and run work you or your workspace start.

As with any hosted service, the people who operate CoDev's production infrastructure could technically reach data stored on it, including workspace disks. We limit that access to the operators who run the service and use it only to keep the service running, respond to security incidents or abuse, meet legal obligations, or help when you ask us to. We do not access your workspace contents, credentials or environment variables for any other purpose. Our staff access policy sets out these limits, the records we keep, and when we tell you about an access; the Security and privacy page lists each way staff could reach your data and how we are closing it. Companies on Teams or Enterprise can instead run CoDev in a cloud account they own, which CoDev staff cannot access unless they grant it.

When something fails in CoDev, we record the error message you saw (or, for a button that stops responding, which kind of control it was), with secrets, emails, links, file paths and names removed, and use Anthropic to tell whether it is a bug. If you choose to report it, our team receives that redacted error, never your code, files or identity.

Who receives information

We use service providers for hosting and execution (including Microsoft Azure), network delivery (Cloudflare), database/storage (Supabase/PostgreSQL) and Redis infrastructure, payment processing (Stripe), transactional email (Resend), and inbound support email forwarding (ImprovMX and the receiving mail service). Google and GitHub receive information needed for sign-in and connected features. AI providers receive the task information described above. The providers involved depend on the features and connections you use; the subprocessor list names each one and what it does.

We may disclose necessary information to comply with law, protect people and the service, investigate abuse, resolve payment disputes, or support a business transfer subject to appropriate confidentiality and legal safeguards. We do not sell personal information or share it for cross-context behavioral advertising.

Cookies and choices

Essential cookies support sign-in, security and your privacy preference. Local browser storage also remembers interface preferences. Optional analytics remains off until you choose “Allow analytics.” You can reject it or change your choice using Privacy choices at any time. Global Privacy Control and Do Not Track signals disable optional analytics. See our cookie notice.

Retention and deletion

We keep account and workspace data for the period needed to provide the service. Deleting a workspace removes its active workspace records and invokes cloud disk/snapshot cleanup. Account deletion removes your profile, stored personal connections, environment variables, private imported conversations and access tokens; shared contributions may remain attributed to “Deleted account.” Billing, security, legal and backup records may remain where necessary. See data retention for the limits and process.

Your privacy rights

Depending on your location and applicable law, you may request access, a copy, correction, deletion, restriction or portability of your information; object to processing; withdraw consent; or appeal a decision. These rights have legal exceptions. We do not penalize you for exercising them. California residents may also have rights to opt out of sale/sharing and limit certain uses of sensitive information; we use account credentials to provide and secure your requested service.

Use Settings → Profile to download your account data and for verified account deletion. The download covers your profile, sign-in methods, environment variable names (never their values), workspaces and memberships, chats and messages you wrote, workspace history you authored (without file contents), security events and billing record; each list holds its most recent 5,000 entries and says when more exist. It leaves out credentials, other members' messages and workspace files, so it is not a complete personal-data export. For a broader request, an authorized-agent request, an appeal, or help accessing your account, email admins@trycodev.com. We verify identity proportionately, respond within applicable legal deadlines, and explain any refusal or permitted extension. You may complain to your local privacy regulator.

International processing and legal bases

Our service providers may process information in countries other than yours, including the United States, where privacy protections differ. Where required, transfers must be supported by applicable legal safeguards. If European or UK data-protection law applies, we rely on contract for requested account/workspace and billing services, legitimate interests for proportionate security, support and service operation, legal obligations for required records, and consent for optional analytics where required. You may withdraw consent without affecting earlier lawful processing.

Security, age and changes

We use access controls, encryption for stored credentials and other technical protections, but no system is guaranteed secure. CoDev accounts are intended for adults aged 18 or older. Contact us if you believe a child has provided personal information. We publish policy changes here and provide additional notice or obtain consent when required for material changes.