ytimesy Privacy Policy

Last updated: 2026-11-04.

This policy explains what ytimesy ("ytimesy", "we") collects when you use ytimesy, why, who else receives it, how long we keep it, and your choices. It is part of our Terms of Service. By creating an account or signing in, you acknowledge this policy.

The short version

We collect what we need to run coding agents on your repositories: your account, the repositories you connect, your chats, the files and commands in your workspaces, and encrypted credentials you give us.
Your prompts and the code an agent reads go to the model provider you chose, under your own key or subscription and your own agreement with that provider.
Training data. The "Help improve ytimesy" setting is shown during onboarding with its switch on (off in the EU, EEA, UK and Switzerland). When it is on, your prompts and messages, your code and repository context, and results can be used to improve ytimesy. We never use AI models' replies or model-generated code as training data. Turn it off at any time in Settings → Privacy. Today no training export exists; see Training Data.
Memory is on by default; you can turn it off or clear it.
We record masked session replays of how the web app is used, to fix bugs and improve the product. Inputs, chat content, code and secrets are masked. No replays are recorded in the EU, EEA, UK or Switzerland.
We do not sell your personal information or use it for advertising.
You can delete chats, projects, your memory, and your whole account yourself.
ytimesy is for people 18 and older.

1. What we collect

Account. Name, email address, profile image, and the identity you sign in with (email and password, GitHub, or Google). Passwords are stored only as hashes. GitHub sign-in asks only for your profile and email (user:email). Sign-in tokens from GitHub or Google are stored encrypted. For each signed-in session we store the IP address and browser user agent. We record which version of the Terms and this policy you accepted, when, and where (sign-up, sign-in, or the in-app update notice).

GitHub connection. If you install the ytimesy GitHub App, we receive the list of accounts and repositories you selected and can act on them. If you connect your GitHub account, we store its access token encrypted. See section 6.

Your content. Prompts and messages, file attachments, the agent's replies and tool output (commands run, their output, file contents read and written, diffs), pull requests the agent opens, project settings, notes agents save about a project, and pull request evidence (screenshots, short screen recordings and command output captured from a pull request's code). If you continue a local agent session in ytimesy, we also store the session transcript you upload; the command-line tool redacts secrets from .env files before uploading.

Workspaces. Each chat runs in a cloud virtual machine with a copy of your repositories, installed dependencies, and anything the agent or you create there. Read-only subagents share their chat's machine; machines are not shared between accounts. Idle workspaces are saved as snapshots so you can resume (section 9).

Previews. Apps an agent runs in your workspace can be opened as previews. Previews are private to you. If you turn on sharing for a preview, anyone with its link can open it until you turn sharing off.

Credentials you give us. Model API keys, sign-in tokens for subscriptions you connect (ChatGPT/Codex, a Claude Code setup token, or xAI), project secrets, and MCP server headers and sign-ins. These are encrypted with application-level encryption before storage and decrypted only to run your work. Project secrets are placed on your workspace machine when your work needs them; model keys and subscription tokens are not.

Memory. Memory is on by default. While it is on, an agent using your connected model writes notes from your chats (preferences, project facts, lessons) into a private memory repository for your account, and we create numeric embeddings of those notes so agents can recall relevant ones. You can view memory, clear it, or turn it off in settings.

Training settings. Whether "Help improve ytimesy" is on, each change with its time, source and the policy version shown, and which projects you excluded.

Usage records. Which models and runtimes you used, token counts, workspace time and machine size, and usage against our 5-hour and weekly limits. We take no payments, so we hold no payment details.

Product analytics. We use PostHog. From the web app we send a fixed set of events (for example "app viewed", "prompt submitted", which inspector panel was opened, page speed measurements), the pages you view (the page's path only, with project pages shown as "/projects"; sign-in pages are not counted), the referring site's host name, time spent and how far you scrolled on a page, and error reports (the error and stack trace with secrets redacted, and the page path). Our servers send events such as "account created", the steps of connecting GitHub and a project, "turn started/finished", "pull request created" and agent performance measurements (timings, token counts, tool outcomes). Events are linked to your account id, or before you sign in to a random analytics identifier, and do not contain prompts, code or repository names. The web app sends them to PostHog through our own domain. IP-based location lookup is turned off. If your browser sends "Do Not Track", the web app sends no analytics or session replays; server events are still recorded.

Referral attribution. When you first visit, the web app keeps where you came from (a ref or utm_* link tag, the referring site's host name and the page you landed on) in a first-party cookie for 90 days, and when you create an account we save it with your account and send it to PostHog with the "account created" event, so we can see which channels bring sign-ups. It is not set for visitors in the EU, EEA, UK or Switzerland, or under "Do Not Track" or Global Privacy Control.

Session replays. While you are signed in, PostHog records masked session replays of how you use the web app (clicks, scrolling, which screens you open), linked to your account id, so we can fix bugs and improve the product. Replays are masked in your browser before anything is sent: all text, including chat content, is masked, every input is masked, and code, diffs, terminals, file contents, secrets, images, previews and the desktop view are blocked. Sign-in pages, network traffic and console logs are not recorded. Replays are kept for 30 days and deleted with your account. We do not record session replays for people in the EU, EEA, UK or Switzerland, based on the country our hosting provider reports for your connection. To object to replays elsewhere, email us (section 11).

Logs and bot checks. Our hosting provider keeps operational logs (requests, errors, timings) that can include IP addresses and identifiers, for security and debugging. Email sign-up, sign-in, password reset and verification resend use Cloudflare Turnstile to block bots.

Email. We send account emails (email verification and password reset) from noreply@ytimesy.com; replies go to support@ytimesy.com. Mail you send to our addresses is forwarded to a mailbox hosted by Google.

Cookies and browser storage. We use a sign-in cookie, a few cookies that remember interface state (such as open panels) and the referral attribution cookie described above. The browser also stores unsent message drafts and the analytics identifier. We use no advertising cookies.

2. Children

ytimesy is not for anyone under 18 (see Terms, section 2). We do not knowingly collect personal information from anyone under 18. If we learn or have reason to believe that an account belongs to someone under 18, we delete the account and its data as described in section 10. If you believe a minor has an account, contact support@ytimesy.com.

3. How we use it

To provide the service: sign you in, connect repositories, run agents and workspaces, save and restore chats, push branches and open pull requests, show your history and evidence.
To call the model you chose with your prompts and the context the agent needs.
To keep your memory, if it is on.
To enforce usage limits and the Terms, prevent abuse, and keep the service secure.
To support you when you contact us.
To understand and improve the product using the analytics described above.
If "Help improve ytimesy" is on, to improve ytimesy's agents and to train or evaluate models, using only the data listed in section 4.

We look at your code, chats or workspaces only (a) when you ask us to (for example in a support request), or (b) to investigate abuse, a security incident or a legal request.

For people in the EU, EEA, UK and Switzerland, see section 11 for our legal bases.

4. Training data

The setting. "Help improve ytimesy" is shown during onboarding with its switch on (off in the EU, EEA, UK and Switzerland), and your choice is saved when you continue. Until you make a choice there or in Settings, it is off. You can change it at any time in Settings → Privacy, and you can exclude a single project in its settings.
What can be used when it is on, from chats that run while it is on and are not in an excluded project:
- your prompts and messages;
- your code and repository context (files and command output as they were before the agent changed them);
- results: command and test output, exit codes, which files changed, pull request outcomes, and feedback you give on a turn.
What is never used, whatever the setting:
- AI models' replies and model-generated code, from any provider;
- attachments, your memory, and imported local agent sessions;
- credentials stored in ytimesy (model keys, subscription tokens, GitHub tokens, project secrets, MCP sign-ins), which are kept outside chats.
A secret that appears inside a chat (for example pasted into a message or printed by a command) is part of that chat. ytimesy does not yet remove such secrets automatically, so keep the setting off, or exclude the project, for code with secrets or code you are not allowed to share.
When it is off, nothing from your chats is used. Turning it off marks every earlier chat of yours as not eligible at once.
Today, eligible chats stay in our main database like any other chat. Nothing has been copied out for training, and no model has been trained on them. We will update this policy before we copy eligible data for training.
We do not sell training data and do not give it to model providers or others for their own use.

5. Model providers and subscriptions

Agents run on the model you pick. When an agent runs, ytimesy sends your prompt, conversation history, attachments, and the code, files and tool output the agent needs to the provider you chose:

Your API key or endpoint (OpenAI, Anthropic, Google Gemini, OpenRouter, or an OpenAI-compatible endpoint you set): the provider processes the request under your agreement with it, including its retention and training policies. With OpenRouter, the request also goes to the model host OpenRouter routes it to. Web search, where available, is the provider's own search.
Your subscription (ChatGPT/Codex through ChatGPT sign-in, Claude through a Claude Code setup token, or Grok through xAI sign-in, where offered): requests are sent to that provider on your account, under your agreement and settings with it. ChatGPT requests can pass through relay servers that run our code on their way to OpenAI.

Your keys and sign-in tokens are stored encrypted and used by our servers; they are not placed on your workspace machine. For a Claude subscription, the machine holds only a short-lived grant for the current turn, and our server swaps in your token when forwarding to Anthropic.

We do not control how these providers handle data. Check your provider's settings if you want them not to retain or train on your data. We do not have zero-data-retention agreements with model providers, because the agreement for your traffic is yours, not ours.

ytimesy itself uses Cloudflare Workers AI to create embeddings of your memory notes.

6. GitHub

The ytimesy GitHub App works only on the accounts and repositories you choose when you install it. Our App configuration asks for: repository contents (read and write), pull requests (read and write), checks (read), issues (read), actions (read) and metadata (read). GitHub shows the exact permissions before you install. We use this to clone your code into your workspace, push branches, open and update pull requests, read check results and CI job logs, and read comments that mention the ytimesy GitHub App on pull requests ytimesy opened, so your chat can fix a failed check or follow a collaborator's request. Agents do not merge pull requests; you can merge one yourself from ytimesy's pull request panel. You can change the App's repositories or uninstall it at any time in GitHub settings.

Pull request descriptions written by the agent may link to evidence pages on ytimesy. Those pages and their files are visible only to you, signed in to ytimesy; other people on the repository cannot open them.

7. Who else receives data (subprocessors)

We use these providers to run ytimesy. Each processes data only to provide its service to us.

Provider	What it does for us	Data	Location
Cloudflare, Inc.	Web app and API hosting (Workers, Durable Objects), private object storage (R2: encrypted chat checkpoints, encrypted attachments, encrypted imported sessions, pull request evidence), memory repositories, memory embeddings (Workers AI), sending email, receiving email, bot checks (Turnstile), network egress, logs	All categories above	Global network; storage location chosen by Cloudflare
Supabase, Inc.	Main database (Postgres): account, projects, chats, usage, encrypted credentials	Account, content, credentials (encrypted), usage	United States (AWS us-east-2, Ohio)
Freestyle	Workspace virtual machines, previews and machine snapshots	Repository copies, workspace files, command output, project secrets on the machine	Chosen by Freestyle
PostHog, Inc.	Product analytics, error reports and masked session replays	Account id, product events, error reports, masked replays	United States
Google LLC	Sign-in with Google (if you use it); mailbox for email sent to our addresses	Account; email you send us	United States
GitHub, Inc.	Sign-in and repository access	Account, repository content	United States

Providers you choose. Model providers you connect (OpenAI, Anthropic, Google, xAI, OpenRouter and its hosts, or your own endpoint) receive prompts, code and tool output for each turn. They act under your agreement with them, not as our subprocessors; we list them for transparency.

We will update this list when we add a subprocessor.

We may also disclose data if required by law, to protect people or the service, or as part of a transfer of the service (with notice to you).

8. Who can see your data in ytimesy

Your projects, chats, memory and evidence are visible only to your account; ytimesy has no team sharing yet. A preview is visible to others only while you share its link. Pull requests and commits the agent pushes are visible to whoever can see the repository on GitHub.

9. How long we keep it

Data	Kept
Account, projects, chats, tool output	Until you delete them or your account
Deleted (archived) chats and projects	30 days in Settings → Archived, where you can restore them or delete them for good; then deleted automatically
Workspace machine	Pauses after a few minutes without activity. After about 1 hour idle, it is saved as a snapshot (with project secrets removed) and the machine is deleted
Saved workspace snapshot	90 days, then deleted; sooner when you delete the chat, project or account
Chat checkpoints, attachments and imported sessions (encrypted)	Until you delete the chat, project or account
Pull request evidence	30 days, or until the pull request is no longer open, whichever is first; sooner when you delete the project or account
Memory repository and embeddings	Until you clear memory or delete your account
Credentials, keys, secrets	Until you remove them or delete your account
Signed-in sessions (with IP address and user agent)	Until you sign out or they expire
Sign-in, verification and device-login codes	Until used or expired
Training setting changes	The current setting until you delete your account; the change history is kept after deletion as proof of consent and withdrawal
Cloud usage ledger (amounts and times, linked to your account id)	Kept after account deletion as our spend record
Session replays	30 days, or until your account is deleted
Analytics events	PostHog's retention for our project, or until your account is deleted
Referral attribution	The browser cookie for 90 days; on your account until it is deleted
Hosting logs	The hosting provider's log retention period
Database backups	Rolled off on our database provider's backup schedule

10. Deleting your data

Chats and projects: deleting one archives it for 30 days (restorable in Settings → Archived); then, or when you delete it for good there, we delete its messages, saved checkpoints, attachments, workspaces and saved workspace snapshots, and those of its subagent chats. Deleting a project also deletes its environment snapshots, imported sessions and pull request evidence. Your GitHub repositories are not touched.
Memory: "Clear memory" deletes your memory repository and its embeddings.
Account: Settings → Delete account. This immediately deletes your account, projects, chats (including archived ones), attachments, checkpoints, imported sessions, pull request evidence, workspaces and snapshots, memory, credentials and training setting, signs you out everywhere, and asks PostHog to delete your analytics person, events and session replays. We keep only the training-setting change history and the cloud usage ledger described in section 9.
Deleting your ytimesy account does not uninstall the GitHub App, revoke your sign-ins at model providers, or delete data your model provider holds. Uninstall or revoke those directly.
Copies in provider backups and logs age out on their normal schedule.

11. Your rights

Depending on where you live (including the EU, UK and California) you may have the right to access, correct, delete, or receive a copy of your data, to object to or restrict some processing, and to withdraw consent. You can delete data and turn off training use and memory yourself in the app. ytimesy has no self-serve data export yet; for a copy of your data or anything else, email support@ytimesy.com. We may need to verify your identity. We will respond within the time the law requires and will not treat you differently for exercising a right. If we deny a request, you can appeal by replying to our answer.

EU, EEA, UK and Switzerland. Where the GDPR, the UK GDPR or the Swiss data protection law applies, our legal bases are:

contract: running the service and your account;
legitimate interests: security, abuse prevention, fixing bugs and improving the product, including product analytics (and, outside these regions, masked session replays);
consent, only where we ask for it, such as the "Help improve ytimesy" setting;
legal obligations.

For people in these regions we do not record session replays, and the "Help improve ytimesy" switch starts off. You have the right to access, correct, delete and receive a copy of your data, to restrict processing, to object to processing (including product analytics), and to withdraw consent at any time. To exercise a right, email support@ytimesy.com with "Privacy Request" in the subject. You can also complain to your local data protection authority. Transfers out of these regions rely on the Standard Contractual Clauses and the UK International Data Transfer Addendum (section 13).

We do not sell or "share" personal information for cross-context behavioural advertising as those terms are defined in California law.

12. Security

Credentials, keys, secrets, chat checkpoints, attachments and imported sessions are encrypted at the application level before storage, in addition to provider encryption at rest. Database tables are not reachable from the browser directly. Each chat runs in its own virtual machine, and machines are not shared between accounts. Connections to ytimesy use HTTPS. No system is perfectly secure; if a breach affects your data we will notify you as the law requires. Report vulnerabilities to support@ytimesy.com.

13. International transfers

Our providers process data in the United States and elsewhere. For transfers of personal data out of the EU, EEA, UK or Switzerland, we rely on the European Commission's Standard Contractual Clauses and, for the UK, the International Data Transfer Addendum, as offered in our providers' data processing terms.

14. Changes

We may update this policy. When we do, we update the date above and show a notice to signed-in users in the app. If you keep using ytimesy after the update, the updated policy applies. A policy change never turns training use on for you; that needs your own action.

Contact

Support, privacy and legal requests: support@ytimesy.com.