# Privacy Policy — caas.sh Last updated: October 12, 2026 coresourceai, Inc. (“coresourceai,” “we,” “us” or “our”) operates caas.sh. This Privacy Policy explains what information we collect when you use our websites, the caas.sh console, CLI and API, and our other services (together, the “Services”), how we use and share it, how we protect it, and the choices you have. It is part of our [Terms of Use](https://caas.sh/terms). In this policy: - **“Personal information”** means information that identifies you or can reasonably be linked to you, such as your name, email address or IP address. - **“Customer content”** means the code, files, images, data and other material that you and your workspace’s members put into the Services, and what your sandboxes, MCP servers and workflows produce. - **“Workspace”** means the account that holds your sandboxes, servers, runs, images, members and balance. ## I. Information We Collect We collect information you give us, information our systems record when you use the Services, and information from the services you choose to connect. ### Information you provide when you register - **Account.** Your email address, a password and a display name. When you sign up with email, your display name starts as the part of your email address before the “@”. We store only a salted, one-way hash of your password, never the password itself. - **Sign-in with Google or GitHub.** If you choose to sign in this way, we receive your name, email address, profile picture address and account ID from that provider, along with the tokens it issues for the sign-in. We ask Google for your basic profile and email address, and GitHub for read access to your profile and email addresses. - **Workspaces and members.** The workspaces you create or join, each member’s role (owner, operator or viewer), and the invitations you send, including the invited person’s email address. - **API keys.** Each key’s name, scopes and expiry, who created it and when, whether it has been revoked, and its last six characters. We store only a hash of the key itself. - **Billing.** You add funds and save cards on Stripe’s own payment pages, so your card number goes to Stripe, not to us; we never receive or store full card numbers. We give Stripe your name and email address to create your customer record. We keep your Stripe customer ID; each saved card’s brand, last four digits and expiry date, which the console shows; each top-up’s amount, result and receipt link; and the notices Stripe sends us about your payments. We also keep your balance and its history, your auto-recharge settings, and the email address for spend alerts. - **Messages.** What you send us when you email us. ### Information we collect through technology - **Usage and metering.** For each workspace we meter what the Services use, such as CPU time, memory, disk, egress, storage, MCP Gateway requests, workflow steps and registry pulls, so that we can bill for it and show it to you. - **Requests and IP addresses.** When you use the Services, we receive your IP address, your browser or client’s user agent, and details of each request. We store your IP address and user agent with each signed-in session, use IP addresses to rate-limit requests and stop abuse, and keep request logs, which can include them, for 7 days. Each request has a request ID that we use to trace problems. - **Audit log.** Each workspace has an audit log of account actions, such as creating or revoking an API key, inviting a member or changing a secret. Each record holds who acted (their user ID and email address), what they did and to what, when, the result and the request ID. It doesn’t record IP addresses or secret values. - **Egress log.** For each sandbox we log its outbound requests: the time, the method, the destination host and path (never the query string), whether the request was allowed and under which rule, whether a credential was added, the response status and the bytes sent. We never store request or response bodies or header values. - **MCP and model traffic.** For each request to a hosted MCP server, we record the method, the tool called, the client token that called it, the status, the timing and the outcome, never the arguments or the response. For each request through the Model Gateway, we record the provider, model, status, token counts, latency and cost, never the prompt or the response. - **Cookies and browser storage.** The console sets a session cookie that keeps you signed in for up to 7 days after you last use it, and a short-lived cookie, lasting 5 minutes, while you sign in with Google or GitHub. It also remembers the last workspace you chose in your browser’s local storage. The CLI keeps your API key in a configuration file on your computer. Neither caas.sh nor the console uses advertising or analytics cookies, tracking pixels or third-party analytics. Our pages load fonts from Google Fonts, so your browser sends your IP address to Google when it fetches them. ### Customer content The Services store and process the customer content you put into them: files in your sandboxes and their snapshots, images in the container registry, workflow files, inputs, outputs and checkpoints and the files carried between steps, the output and logs your workloads write, and metadata about traffic to your MCP servers. We process customer content only to provide the Services to you. We don’t look at it except as needed to operate, secure or support the Services or to comply with the law, and we don’t use it to train machine learning models. If your customer content includes personal information about other people, you are responsible for it, and we process it on your behalf. When someone visits an app you publish at a sandbox URL, we pass the visitor’s IP address to your app. ### Connected credentials and secrets When you connect a credential, such as a GitHub token or a model provider’s API key, or store a secret for your workloads, we encrypt it (with AES-256-GCM) before storing it. We use it only to make the requests you configure: it is added to your sandbox’s outbound requests to the hosts you allow, and your sandbox sees a placeholder instead of the value. ### Children’s privacy The Services are not for children. You must be at least 18 to use them, and we don’t knowingly collect personal information from children under 13, or under 16 where the law sets that age. If you believe a child has given us personal information, contact us at legal@coresource.ai and we will delete it. ## II. How We Use and Share Information ### How we use information - to provide and operate the Services, including running your workloads, metering usage and drawing down your balance; - to create and secure your account, verify your email address and reset your password; - to process payments and send billing notices, such as low-balance alerts, spend alerts and failed auto-recharge notices; - to keep the Services secure, including rate limiting, preventing fraud and abuse, and enforcing our Terms of Use; - to answer your questions, troubleshoot problems and improve the Services; and - to comply with the law. We don’t use your information for advertising. ### How we share information We don’t sell personal information. We share it only as follows: - **Service providers.** Cloudflare hosts the Services and provides their compute, storage, databases, logs and email delivery. Stripe processes payments and sends payment receipts. Workflow runs execute on Constal, a workflow engine that coresourceai, Inc. also operates, which receives each workflow run’s definition, inputs, step logs and outputs and the files carried between steps. Google and GitHub sign you in, if you choose them. Google Fonts serves the fonts on our pages. Each third party provides its service to us under its own terms and privacy policy. - **Your workspace.** Members of a workspace can see each other’s names, email addresses and roles, and the audit log shows who took each action. - **Services you direct us to.** When your sandboxes, servers or workflows send requests to other services, we send them, with the credentials you configure, where you tell us to. - **Legal reasons.** When we believe in good faith that the law, a court order or a government request requires it, or that it is needed to protect the rights, property or safety of coresourceai, our customers or others. - **Business transfers.** If coresourceai is involved in a merger, acquisition, financing or sale of assets, your information may be transferred as part of it, subject to this policy. - **With your consent.** In any other case, only when you ask us to or agree. ### How long we keep information - Account, workspace and billing information: for as long as your account exists, and billing records for as long as tax and accounting law requires. - Signed-in sessions: a session expires after 7 days without use. We delete its record, with its IP address and user agent, when you sign out, or when it is next presented after it has expired. - Request logs: 7 days. - Payment notices from Stripe: 30 days after we process them. - A sandbox’s recent logs, egress log, MCP traces and Model Gateway records: 7 days. Its log archives, which include the egress log: 90 days. Its lifecycle events: 30 days. - The audit log: 365 days, or a workspace’s newest 1,000,000 records if that is fewer. Records keep the email address of the member who acted after that member leaves the workspace. - Files carried between workflow steps and workflow checkpoint images: 7 days after the run ends. Uploaded source archives: 7 days. - Revoked or expired MCP access tokens: 30 days. - Customer content: until you delete it or your account is deleted. Deleting a sandbox deletes its snapshots, log archives, traces and egress log. When we delete your account at your request, we delete your personal information and customer content, except what we must keep by law and the records we keep for billing, tax and security, such as the audit log. Copies in logs expire on the schedules above. ## III. How We Protect Information We use technical and organizational measures designed to protect your information. The Services are served over HTTPS, and the session cookie is sent only over HTTPS and can’t be read by scripts. We store passwords, API keys and MCP client tokens only as one-way hashes, and we encrypt connected credentials and secrets. Each sandbox runs in its own isolated environment, and it can reach only the outbound hosts its policy allows. No system is perfectly secure, so we can’t guarantee the security of your information. If you think your account has been compromised, contact us at legal@coresource.ai. coresourceai is based in the United States. The Services run on Cloudflare’s global network, so your information is processed in the United States and may be processed in other countries where our service providers operate. Where the law requires, we protect such transfers with appropriate safeguards. ## IV. Your Rights Regarding Your Personal Information You can see much of your information in the console, where you can also revoke API keys, remove connected credentials and, as a workspace’s owner, change members’ roles or remove them. To ask for access to, correction, deletion or a copy (export) of your personal information, or to delete your account, email us at legal@coresource.ai. We will verify your request and respond within the time the law requires. Depending on where you live, you may have further rights. California residents have the right to know what personal information we collect, use and disclose, to access, correct and delete it, and not to be discriminated against for exercising these rights; because we don’t sell or share personal information for targeted advertising, there is nothing to opt out of. If you are in the European Economic Area, the United Kingdom or Switzerland, you may also object to or restrict our processing, ask to move your information, and complain to your local data protection authority. We process personal information to perform our contract with you, for our legitimate interests in running and securing the Services, to meet legal obligations, and with your consent where we ask for it. You may use an authorized agent to make a request on your behalf. If your request is about personal information in another customer’s content, contact that customer, who controls it. We will help them respond. ## V. Links to Other Websites The Services link to and work with websites and services we don’t control, such as GitHub, Google and Stripe. This policy doesn’t cover them. Their own privacy policies apply, and we encourage you to read them. ## VI. Changes to Our Privacy Policy We may update this Privacy Policy. When we do, we will post the new version at caas.sh/privacy and change the date at the top of this page. If a change is material, we will also tell you before it takes effect, by email or in the console. ## VII. Contact Us Questions or requests about this Privacy Policy or your personal information? Write to us at legal@coresource.ai, or by mail at coresourceai, Inc., 2261 Market Street STE 68771, San Francisco, CA 94114, US.