Privacy Policy

1. Who we are

Technically Not is made by Technically Not LLC, a Colorado limited liability company in Pueblo, Colorado, USA (“we”, “us”). We are the data controller for the limited personal data described here that reaches us.

2. What the app does with data

Your projects stay on your PC

Technically Not reads and writes your project folder locally. It doesn’t upload your files, code, prompts or terminal output to us, and our servers never receive them.

What the app stores on your PC

The app keeps its own data in %AppData%\TechnicallyNot.ai on your computer. This includes:

• Settings, the last project folder, and agent setup (limits, custom agents, rules).
• Terminal session records and scrollback, so work survives a crash or restart.
• Files you attach in the composer. These are deleted automatically after 30 days.
• Per-project agent context notes and Git worktrees made for agents.
• The voice engine and speech model, once you set up voice prompts. Your recordings aren’t kept.
• If you set up Alerts to Discord, your own webhook address, encrypted by your operating system’s secure store.
• Nothing about a Technically Not account. If an older version saved a sign-in on this computer, the current version deletes that file when it starts, without reading it.

None of this is sent to us. Uninstalling the app and deleting that folder removes it.

Agent sign-ins

Technically Not never asks for or stores passwords or sign-in tokens for AI agents. Each agent CLI (for example Claude Code, Codex or Cursor Agent) signs in by itself. To show which account and plan a panel uses, the app reads the account email and plan name (for example Pro or Max) from the CLI’s own sign-in file on your PC. The token in that file is never copied, shown or sent to us. The one use the app makes of it is for Claude Code, to ask Anthropic how much of your plan is used (see When the app goes online). If you add your own Claude Code or Codex API key instead, it’s kept encrypted on your computer by your operating system’s secure store and never sent to us.

When the app goes online

• Agents you run. An agent sends your prompts and the files it works on to its own provider, under that provider’s terms and privacy policy. Technically Not doesn’t send that data anywhere else.
• App updates. The app looks for updates on GitHub, in the releases of its public source code, over HTTPS at start and every 4 hours, and offers any new version in the app; it downloads and installs only when you choose to (or when you turn on automatic downloads in Settings › Updates). The request carries no sign-in and no account. GitHub sees your IP address, as with any download, under GitHub’s own privacy policy. We don’t receive it.
• Agent CLI version checks. For agent CLIs published on npm, the app asks the public npm registry for the latest version number, shortly after start and every 6 hours. Only the package name is sent.
• Voice setup. The first time you set up voice prompts, the app downloads the speech engine from GitHub and the speech model from Hugging Face, once (another model, for another language or better accuracy, downloads the first time you pick it). After that, speech is turned into text on your PC; your audio is never sent anywhere.
• Your Claude plan’s usage. To show how much of your Claude plan is used (Settings › Usage), the app asks Anthropic’s own server, with the sign-in Claude Code already keeps on your PC. That sign-in goes to Anthropic and nowhere else. Codex usage is read from Codex’s own files on your PC, with no request.
• Your own relay. If you connect several computers on the Devices page, the app talks to a relay you run yourself, on your own Cloudflare account, at the address you enter. Messages are sealed on your computers and the relay cannot read them; it stores no message and remembers only which devices connected and when. We run no relay and receive nothing from yours. What Cloudflare sees and charges is between you and Cloudflare.
• Alerts to Discord (off by default). If you paste your own Discord webhook in Settings › Agents, the app sends a short message to it when an agent needs you, finishes, or seems stuck. It goes to Discord and nobody else.
• Git. The app runs git on your computer for Own worktree (a branch and folder per agent). Agents you run may also use git; anything they send to your repository’s remote uses your own Git credentials.

The app also runs small servers bound to 127.0.0.1 (your own PC only) so agents can report live status. They can’t be reached from the internet.

No telemetry

The app sends no analytics, usage data or crash reports to us or anyone else.

Community total (no longer in the app)

The current app has no community total and sends nothing for it. Versions up to 0.3.68 had a setting, off by default, “Add my agent work to the community total”. While it was on, and you were signed in, the app sent our accounts server one number per day for that computer: how many tokens Claude Code and Codex used that day, read from their own logs on your PC. With it we stored the day, a random device ID the app created, and a link to your account, so the rows can be deleted with it. No names, prompts, code, file names or project details were sent. Deleting your account removes your rows, and rows older than 40 days are removed whenever a new count arrives. What is left goes when the accounts server is switched off (What happens to this data).

3. What this website collects

This website is a set of static pages with no sign-in. It sets no cookies, runs no advertising trackers, and loads no third-party fonts. That’s why there is no cookie banner. We count visits with Cloudflare Web Analytics, which uses no cookies and no fingerprinting and doesn’t identify you: it records the page, the site that linked you here, your browser type and country, as totals. Apart from that Cloudflare counter, no page loads a third-party script. One more request goes to another company: to show the star count on the GitHub buttons, your browser asks GitHub (api.github.com) for the project's public details, keeps the number in its own storage (local storage, not a cookie) and does not ask again for an hour, or for ten minutes when GitHub gave no answer; GitHub sees your IP address and browser type, as with any request, under GitHub's own privacy policy, and we receive nothing from it.

Our web host, Cloudflare, keeps standard server logs (such as IP address, browser and pages requested) for security and to keep the site running. We don’t use them to identify you.

One exception, for accounts made earlier: a link in an email we already sent (to confirm a new account or to reset a password) still opens a page here that finishes that step with our accounts server. The server may then set a single sign-in session cookie. It is strictly necessary for that step and isn’t used for tracking. Those pages carry no visit counter, and nothing on this website links to them.

Which link brought you here

Our links on listings and social media carry a source name, such as ?ref=tiktok (or utm_source). When you arrive through one, your browser remembers the first and the latest source name, with the time, in its own storage (local storage, not a cookie). Once per browser session it also asks our accounts server to add one to a daily count for that source and page. That count is only a number per day, source and page: no IP address, browser details, account or anything else that identifies you is stored with it. To stop abuse, the server keeps a keyed hash of your IP address for up to a day to limit how often it counts. This website has no sign-up any more, so the remembered names stay in your browser; only an older sign-up finished from an emailed link (above) saves them with that account. Clearing this site’s data in your browser removes the remembered names.

TikTok ads

A link in one of our TikTok ads carries a TikTok click ID (ttclid). When you arrive through one, your browser keeps that ID in its own storage (local storage, not a cookie) for up to 28 days. No TikTok code runs on this website, and nothing is sent to TikTok when you visit, download the app or use it. The ID was used to tell TikTok about new accounts and purchases, and this website now has neither. One case is left: if you finish an older sign-up from an emailed link (above) in a browser that holds such an ID, our accounts server sends TikTok the click ID, that an account was made, and one-way SHA-256 hashes of your email address and account ID, never the address itself. For accounts made earlier we keep the click ID on the account and a record of each report (its event, time and whether TikTok accepted it). TikTok handles what it receives under its own privacy policy. To remove the ID, clear this site’s data in your browser, or email support@technicallynot.ai and we’ll remove it from your account.

Email list

This website has no sign-up form for the email list any more. If you joined the list earlier, we store the address, when you signed up, the source name above (if any), and the dates you confirmed or unsubscribed. Nothing is sent to an address that was never confirmed, and such an address is deleted when the accounts server is switched off. People on the list get the build videos. No spam, and we don’t share the list. Every email has an unsubscribe link (it opens a page where one button unsubscribes you), and support@technicallynot.ai works too. After you unsubscribe we keep the address with the unsubscribe date only so we never email it again; ask us and we delete it completely. The emails are sent through Resend.

4. Accounts and plans made earlier

Up to version 0.3.68 the app needed a Technically Not account, and a plan could be bought. Neither exists any more: the current app has no sign-in, and this website has no sign-in page, no account page and nothing to buy. This section is about what is left from that time.

What our accounts server still holds

If you made an account, our accounts server at api.technicallynot.ai stores:

• Your account email, and your name if your sign-in provider shared it.
• How you signed in (Google, GitHub where offered, or email and password).
• If you used a password, a hash of it. We never stored the password itself.
• If you turned on two-step sign-in, its secret (encrypted) and hashes of your recovery codes.
• Your plan and subscription status.
• Records of your sign-in sessions.
• For each computer you signed in to the app on: its device name, a random device ID the app created, and when it was last used.
• Short-lived counters that limit repeated sign-in attempts. They hold a keyed hash of your IP address or email, never the raw value, and are cleared within a day.
• If you connected Discord on the account page: your Discord user ID, username and when you connected it, which our Discord bot used to add you to the Technically Not Discord server and to give or remove the Pro and Founder roles. Discord’s sign-in token was used once and never stored. Deleting your account removes the link and those roles.
• If a Founder Yearly plan ended without renewing (including when the account was deleted while on it), a one-way hash of your sign-in identities. It stays after the account is deleted, but only until the accounts server is switched off; then it is deleted with the rest.
• The source names and the TikTok click ID described in section 3, if your browser held them when the account was made.

Stripe holds payment details; we don’t. Your AI provider keys and your project files were never sent to us.

Older versions of the app still use that server

A copy of the app up to 0.3.68 still signs in to that server and checks your plan there. It also asks that server for updates (the server sees your IP address and app version) and for the list of Claude models, and sends the community total if you turned it on. Updating to the current version ends all of this: it deletes the saved sign-in from your computer and never contacts that server.

What happens to this data

We keep the accounts server running for one reason: so installed older copies can still sign in and fetch the update that moves them to the free version. We switch it off three months after that update is released. On that day we delete all account data on it: the accounts with their sign-in sessions, devices and Discord links, the Founder hashes, the community total counts, and email-list addresses that were never confirmed. The one thing we keep is purchase records that tax law requires us to keep (see How long data is kept).

You don’t have to wait. This website no longer has an account page, so write to support@technicallynot.ai from the account’s email address and we’ll delete the account. An older copy of the app is then signed out; the current version needs no account, and you can get it on the Download page.

Payments

Technically Not is free, and this website takes no payment. A plan bought before the app became free was processed by Stripe (Stripe Managed Payments), acting as Merchant of Record. Stripe collected the buyer’s name, email, billing address and payment details, and handled tax. We never saw full card details. Stripe shared with us what we needed to run and support that subscription: name, email, country, the plan and its status. Stripe’s privacy policy also applies.

5. Support email

If you email us, we use your message and address only to answer you and to keep a record of the conversation.

6. Third parties

• AI agent providers you choose to run (for example Anthropic, OpenAI, Cursor). You have your own relationship with them.
• GitHub hosts the app’s releases, which the app checks and downloads directly, and the speech engine for voice prompts. Links to the source code on this website open GitHub, where GitHub’s own privacy policy applies.
• npm registry answers CLI version checks.
• Hugging Face hosts the speech model the app downloads once for voice prompts.
• Cloudflare for this website. A relay you run yourself is on your own Cloudflare account, under your own agreement with Cloudflare.
• Discord, if you set up Alerts to Discord in the app, or connected Discord to an account made earlier.
• Google and GitHub, only if you used one to sign in to an account made earlier.
• Stripe for plans bought before the app became free.
• Resend sends emails for accounts made earlier and email-list emails.
• TikTok gets the ad report described in TikTok ads, only in the one case named there.

We don’t sell or rent personal data. The only thing we share for advertising is that TikTok ad report.

7. How long data is kept

• Data in the app stays on your PC until you delete it. Attachments are removed after 30 days.
• Account data on our server, for accounts made earlier: until you ask us to delete the account, or until the accounts server is switched off (What happens to this data). Deleting an account removes its email, name, sign-in methods, password hash, two-step sign-in data, sessions and devices; the Founder hash described above stays until the accounts server is switched off, and a canceled subscription record is kept as below.
• Purchase and subscription records from before the app became free: as long as tax and accounting law requires.
• Support emails: up to 2 years after the conversation ends.
• Email list: until you unsubscribe (then only the address and unsubscribe date, as described above) or ask us to delete it. Daily source counts contain nothing personal and are kept as totals.
• Community total daily counts from older versions: removed with the account, and otherwise as described under Community total.
• TikTok ad click ID and report records: until the account is deleted (a record of a purchase report is kept with the purchase record).

8. Your rights

Depending on where you live, you can ask us to access, correct, delete or export the personal data we hold about you, or to object to or limit how we use it. Email us and we’ll answer within one month.

• EU and UK (GDPR). The grounds we rely on are these. An account made earlier: performing our agreement with you, and, now that no plan is sold, our legitimate interest in letting older copies of the app sign in and update until the accounts server is switched off. Purchase records: that agreement and our legal duties. The email list: your consent, which you gave by confirming your address and can take back at any time with the unsubscribe link. The daily count of which link brought you here, with the hashed IP address kept for up to a day: our legitimate interest in knowing which of our links are used and in stopping abuse. The TikTok ad report: our legitimate interest in knowing whether an ad worked. Support emails: our legitimate interest in helping you. You can object to any use that rests on a legitimate interest, and you can complain to your local data protection authority.
• Where data is held. Technically Not LLC is in the United States. Account data, the email list and the link counts are held for us by Cloudflare, emails are sent by Resend, and purchase records are held by Stripe. These companies may process data outside your country.
• California (CCPA/CPRA). You have the right to know, delete and correct your data, and not to be discriminated against for using these rights. We don’t sell or share personal information as those laws define it.

Most app data never reaches us, so you control it directly on your PC.

9. Children

Technically Not is a developer tool and isn’t directed at children under 16. We don’t knowingly collect their data. If you think a child has sent us data, contact us and we’ll delete it.

10. Changes to this policy

If we change how data is handled, we’ll update this page and the “Last updated” date. For important changes we’ll also say so in the app’s release notes, and by email to people on the email list.

11. Contact

Privacy questions or requests: support@technicallynot.ai.