Legal

Privacy Policy

Last updated November 3, 2026

Who we are

Orbit is a product of Tunl Inc., doing business as Aesthetic (“we”, “us”). Our address is 300 Spectrum Center Dr, Ste 400, Irvine, CA 92618. For anything in this policy, email hi@ljnorth.com.

This policy covers the Orbit website and app, the Orbit connector your AI agent uses, and the orbit-connect command-line tool. It doesn't cover Claude, Codex, Cursor, VS Code, Instagram, YouTube or any other service you use alongside Orbit. Those have their own policies.

What we collect

When you sign in with Google

We ask Google only for the basic sign-in permissions (openid, email, profile). We store your email address(only if Google says it's verified), your name, a handle we make from your email address, and when your account was created and last used. We don't store your Google profile photo, and we can't see your Gmail, Drive, contacts or password.

Orbit's use of information received from Google APIs adheres to the Google API Services User Data Policy, including the Limited Use requirements.

When you connect an agent or app

You can connect Claude, ChatGPT, Codex, Cursor, VS Code or a terminal to Orbit. For each connection we store the app's name as the app reports it, the address it returns to after sign-in, the kind of agent, your computer's name if the orbit-connect tool reports it, and when it was connected and last used. The keys that let your agent talk to Orbit, including your private plugin link, are stored only in scrambled (hashed) form, so we can check them but can't read them back.

What your agent asks Orbit

When your agent uses the connector, it sends Orbit requests such as a search for skills or tools that fit what you're building. We use those requests to answer them. Our server logs record only which tool was used, whether it worked, how long it took and how many results it returned, never what you searched for. We don't send these requests to outside AI providers. If that ever changes, we'll update this policy first.

Your agent setup, when you run orbit-connect

The tool reads your setup on your computer, shows you a summary of what it would send, and uploads only after you answer yes. What it can send:

• The names of the skills, plugins, MCP servers and outside services or API hosts you use, with simple counts (for example, how often each was used in the last 30 days and in how many projects).
• For skills you wrote yourself: the skill's description and text, with anything that looks like a secret removed first. This stays private unless you share it.
• How many CLAUDE.md instruction files you have. Only the count is sent, not their names or contents.
• If you have a local AI model (Ollama) installed: a short summary of what you're working on, made by that model on your computer. It includes a project name, a one-line summary, goals, the kinds of help you need, session and commit counts, and a list of numbers (an “embedding”) used to match you with skills. You can skip this with --no-intents.

It never sends secret values, environment variables, command lines or file paths. Our server checks everything again and throws away anything that still looks like a secret.

Which skills you use

If you install Orbit's Claude Code plugin, or let orbit-connect sync --codex add hooks to Codex, a small script on your computer reports when one of your skills runs. It sends exactly four things: the skill's name, whether it was started by the agent or by a /command, whether it worked, and the time. Our server rejects anything else. Your prompts, files, paths, commands and tool output never leave your computer, and a prompt that doesn't start with a skill command makes no network request at all. To turn it off, run claude plugin disable orbit in Claude Code, or remove the Orbit entries from ~/.codex/hooks.json.

Things you do on Orbit

The skills and agents you add, save or install, the people you follow on Orbit, what you tell Orbit you're building (on the site or in a chat with the Orbit connector), your auto-setup settings, and the choices you make about what's public or private.

Sending to your agent and Run it

When you tap Send to my agent or Run it, we record which skill or agent you sent, its version and a fingerprint (hash) of its files, and whether your agent installed it, so it arrives exactly once and unchanged. Run it then opens your own agent (for example Claude Code, Codex or Cursor) on your computer with a prompt we wrote, which you can read before you press Enter. We don't see what you do in your agent afterwards.

When you buy something

Payments are handled by Stripe. Stripe collects your card details; we never see or store your card number. We keep your Stripe customer ID, what you bought or subscribed to, the price, and whether it's active, so we can give you access. If you sell on Orbit, we also record each sale and what you've earned.

Technical data and cookies

Like most websites, our hosting provider records technical details of each request, such as your IP address, browser type and the page requested. We use IP addresses briefly, in memory, to stop abuse with rate limits. We set one sign-in cookie that keeps you signed in for up to 30 days, plus two short-lived cookies during Google sign-in that protect it from forgery. We don't use advertising cookies or cross-site trackers.

Product analytics Planned We may add PostHog to understand how people use Orbit (for example, which pages are visited). If we do, we'll list it here before it's turned on and won't send it your prompts, code or scan contents.

The local scanner (orbit-connect scan)

The scanner turns your agent history into a short summary of what you're building. Here's exactly how it works.

• It reads, on your computer only. It reads your Claude Code and Codex session files on your machine without changing them.
• Your own agent writes the summary. To group your work into projects, the scanner asks the agent you already use (Claude Code or Codex) on your computer, using your own account with that provider, or it uses simple rules if you prefer (--summarizer heuristic). That step happens between you and your agent's provider, under their terms. Orbit doesn't receive it.
• You see it first. The scanner shows you the summary and saves it as a draft file you can edit. Nothing is uploaded until you choose to upload it.
• What can be uploaded: project and build names, short descriptions, the kind of build (an app, a website, a video and so on), short notes on how you work on each, folder names without their paths (for example my-app, never /Users/you/my-app), numbers (tokens, sessions, active days, tool calls, commits in the last 30 days, first and last active dates, how many videos, PDFs and images you made), and the names of the skills and services you used with counts. If you pick a build for it, a small cover image (a JPEG of 200 KB or less) made on your computer.
• What is never uploaded: your prompts, code, file contents, .env values and full file paths. Our server rejects any upload with an unexpected field or with text that looks like a file path or a secret.
• It arrives private and pending. Uploaded builds are visible only to you until you review them at /scan-review. There you can rename, edit, hide or merge builds, choose for each one Public (anyone on Orbit), Stealth (shown without names) or Private (only you), approve them, or discard the scan, which deletes everything you haven't approved.
• Sorting on our server. To sort uploaded builds into kinds and areas, our server may send a build's name, description, workflows and the names of the skills and services it uses to TypeSafe, an AI labeling service that returns only labels. Nothing else from your computer is involved, and TypeSafe processes it only on our behalf.
• Optional background refresh. If you turn on the scanner's hooks, it refreshes at most every 6 hours. A refresh only updates the numbers for builds you already approved and adds any new builds as private and pending. It never runs your agent and never changes text you approved. To stop it, remove the hooks.
• Deleting approved builds. You can make any approved build private at any time. To delete approved builds, email hi@ljnorth.com and we'll delete them within 30 days .

Instagram “tag to add”

You can add a skill to your Orbit by commenting on a creator's Instagram post. We follow Meta's Platform Terms and Developer Policies for everything here.

• Linking your account. You link your Instagram account by sending a one-time code in a direct message to Orbit's Instagram account (@orbit.skills). We store your Instagram-scoped user ID and username and connect them to your Orbit account. We don't get your Instagram password and can't post as you.
• Tagging. When you tag @orbit.skills in a comment on a creator's post, Meta sends us that comment: its text, your username and Instagram-scoped ID, and the post's ID, shortcode and link. We match the post to a skill, add it to your Orbit, and may reply publicly to your comment (for example, to confirm it was added). Anyone who can see the post can see that reply.
• If you haven't linked. Meta sends us comments that tag Orbit even from people who haven't linked an account. We use them only to answer that comment and to let you claim the skill if you link later, and they follow the same retention limits below.
• Retention. We delete or redact the text of comments and direct messages, and the sender details that come with them, 90 days after we receive them. The skills added to your Orbit stay until you remove them. We keep a scrambled (hashed) form of Instagram IDs so that deleted data can't be processed again if Meta resends it.
• Unlinking. You can unlink your Instagram account at any time in your Orbit settings, or by emailing hi@ljnorth.com. We then stop acting on your comments and delete your stored Instagram ID, username and comment and message content within 30 days, keeping only the hashed ID described above.
• No selling. We never sell Instagram data, use it for ads, or share it with data brokers.

Creators and the public catalog

Orbit shows a curated catalog of AI skills and the public posts and videos where creators share them. We collect public information only: public skill files (for example, from public GitHub repositories), and public Instagram and YouTube posts with the creator's handle or channel name, profile picture and public counts. We always credit the creator and link to the original. Instagram posts link out to Instagram, where we keep only a preview image. YouTube videos play in YouTube's own embedded player and are never copied to our servers.

To build the catalog we use AI services (Anthropic's Claude, Groq, Mistral and TypeSafe) to describe, sort and check public content, and data services (Firecrawl, HikerAPI and Apify) to read public web pages and public posts. For the catalog we send them public content only, never data about our users.

Some creators offer a resource to anyone who asks, for example by commenting a keyword or sending a message. Orbit may ask for these the same way any member of the public can, using Orbit's own accounts, and add them to the catalog with credit to the creator.

YouTube. Orbit uses YouTube API Services. By watching YouTube videos on Orbit you agree to the YouTube Terms of Service, and Google's use of data is covered by the Google Privacy Policy. The embedded player is run by YouTube and may set its own cookies. Orbit never asks for access to your YouTube account. If you've given any app access to your Google account, you can remove it at Google's security settings.

Creators: opt out or take down. If you're a creator and want your content, name or picture removed from Orbit, or corrected, email hi@ljnorth.com with the link. We'll remove it within 5 business days and keep it out of future catalog updates.

How we use data

• To run Orbit: sign you in, connect your agents, find skills and tools that fit what you're building, and show you and the people you choose what you've built.
• To count which skills get used, so we can rank skills and show creators how their skills are doing. Creators see totals, never who you are.
• To keep Orbit safe: rate limits, stopping abuse and scanning catalog skills for risky instructions.
• To fix bugs and improve Orbit, and to email you about your account or important changes. We don't send marketing email without your permission.

If you're in the European Economic Area or the UK: we process your data to provide the service you asked for (contract), to keep it safe and working (our legitimate interests), and, for the scanner, usage counts and Instagram linking, because you chose to turn them on (consent, which you can withdraw at any time).

Who we share it with

We don't sell personal information and we don't “share” it for cross-context advertising, as California law defines those terms. We use these service providers to run Orbit, and they may handle data only on our behalf:

What you choose to make public on Orbit is visible to other people. We may also share information if the law requires it, to protect people's safety or our rights, or as part of a merger or sale of the company, in which case this policy keeps applying to your data.

Our servers are in the United States. If you use Orbit from elsewhere, your data is transferred to and stored in the US, relying on safeguards such as the European Commission's standard contractual clauses where the law requires them.

How long we keep it

Your controls

• Disconnect any agent or app on the Connect page. Its key stops working right away.
• Revoke your plugin link by getting a new one on the Connect page. The old one stops working.
• Review, edit, hide, merge, make private or discard your scan at /scan-review.
• Say no: orbit-connect uploads nothing unless you answer yes, and --dry-run shows what it would send without sending it.
• Turn off usage counting with claude plugin disable orbit, or by removing the Orbit hooks from Codex.
• Delete a build or your account yourself in Orbit, under Your data (see below).

Deleting your data

You can delete any build, or your whole Orbit account, yourself under Your data in Orbit. Deleting your account immediately stops every connected agent and app, then deletes your account, connections, uploaded setup and scans, builds, usage counts, suggestions and any linked Instagram data. If you can't sign in, email hi@ljnorth.com from the address you sign in with and we'll do it within 30 days. Deleted data leaves our database backups within 30 days.

From Instagram When live: you can also remove Orbit in Instagram (Settings, then Apps and websites). Instagram then sends us a deletion request, and we delete the Instagram data we hold for you within 30 days and give you a confirmation code and a link where you can check the status.

Deleting your account doesn't delete skill files you already installed on your computer. You can delete those yourself.

Your rights

Depending on where you live, you may have the right to see the personal data we hold about you, get a copy, correct it, delete it, object to or limit how we use it, and withdraw consent. California residents have the right to know, delete, correct and opt out of sale or sharing (we do neither), and we won't treat you differently for using these rights. To use any of them, email hi@ljnorth.com. We'll answer within 30 days, and may need to confirm it's really you. If you're in the EEA or UK, you can also complain to your local data protection authority.

Security

We use HTTPS everywhere, store keys and sign-in codes only in hashed form, check every upload strictly, and limit who on our team can access user data. No system is perfectly secure, so if we learn of a breach that affects your data, we'll tell you as the law requires.

Age

Orbit isn't for children. You must be at least 13 to use it, or 16 if you live in the European Economic Area or the UK. If you believe a child has given us personal data, email hi@ljnorth.com and we'll delete it.

Changes to this policy

If we change this policy, we'll update the date at the top. If a change is significant, for example a new kind of data or a new way of sharing it, we'll tell you by email or on Orbit before it takes effect.

Contact

Tunl Inc., 300 Spectrum Center Dr, Ste 400, Irvine, CA 92618. Email hi@ljnorth.com. See also our Terms of Service.