Privacy policy
Effective November 12, 2026
This policy covers amicklabs.com and Studio (studio.amicklabs.com). Both belong to Amick Labs LLC, an Indiana limited liability company. Questions go to hello@amicklabs.com.
We don't sell your information and we don't show ads.
amicklabs.com
It's a plain website. It has no forms, no accounts, no analytics and no cookies. Vercel hosts it, and like most hosts it may record technical details of each request, such as your IP address.
If you email us, we keep your email.
What Studio collects
- Your account. Your name and email address, from Google, through Clerk. Your account also records which clients you belong to. Anyone with a Google account can create a Studio account, but it shows nothing until we add it to a client.
- Your client details. The client's name and contact email, your weekly price, and any notes we keep about the client.
- What you upload. Raw footage and thumbnails, plus each file's name, type, size and modified time, the order of your clips and any note you add.
- What you do in Studio. Your notes on videos and the words around each note, your approvals, your choices of title, description and thumbnail, and who did each of these.
- Your videos. Each cut, with a transcript of the words spoken in it.
- Billing. Your subscription status and dates, and the Stripe ids that link you to Stripe. Stripe holds your card. We never see your full card number.
- Agent keys. A record of any keys you create for your own agents, and the address of any Grok Bot you connect.
- Your YouTube connection. If you connect a YouTube channel, its id, name and handle, the permissions you granted, and an encrypted Google token. See YouTube below.
- Your Instagram connection. If you connect an Instagram account, its id, username and account type, the permissions you granted, an encrypted Meta token, and a record of each Reel Studio posts there. See Instagram below.
We use all of this to run Studio, make your videos, bill you and talk to you about the work.
Who handles it
| Service | What it does | What it gets |
|---|---|---|
| Clerk | Sign-in with Google, and emailed invitations | Your name, email address and sign-in activity |
| Convex | Studio's database | Everything in the list above, except the video files and your card |
| Stripe | Payments | Your card details, the client name, the email of whoever subscribes, and any billing address you add |
| Vercel | Hosts both websites | Technical details of each request |
| Cloudflare R2 | File storage | Uploaded raw footage, thumbnails and rendered cuts |
| Anthropic (Claude) and OpenAI | The AI agents that help make your videos. We may use either | The transcript, plan and script, your notes, titles and descriptions, and sometimes stills from the video |
| OpenAI | Generates thumbnails | Text prompts, which can describe your video |
| ElevenLabs | Transcribes speech and makes sound effects | The audio of your recordings, words to listen for such as names, and text prompts for sound effects |
| fal.ai | Generates images and video clips | Text prompts, and sometimes still images |
| Grok Bot, from xAI and Cursor | Nathan Amick's AI assistant for managing the work | Client notes, client and video names, titles, version and upload details |
| Google (YouTube) | Publishes to your channel, if you connect it | The finished video, thumbnail, title, description, chapters and captions |
| Meta (Instagram) | Posts Reels to your Instagram account, if you connect it | The finished video, its cover image and caption |
| GitHub | Tracks our work | A client name or video title in an issue or comment, but no video or audio |
These companies have their own privacy policies. They may process data outside your state or country.
How your footage is used
We make videos with our own production system and AI agents. This is how it works.
The system runs on Amick Labs computers on a private network, not on the public internet. It connects to Studio and downloads your raw clips and notes, so copies of your footage sit on those computers.
Your recording's audio goes to ElevenLabs to be transcribed. AI agents then work on the cut, running on Anthropic's Claude, OpenAI's models or both. They can read the transcript, plan and script, your notes, titles and descriptions, and they sometimes look at stills from the video. When we make thumbnails, OpenAI gets a text prompt, which can describe your video. When a video needs generated images or clips, we use fal.ai, which gets the prompt and any still image we give it as a reference. fal.ai may pass these to the company whose model we choose, such as MiniMax, Wan, Google or Grok Imagine. We sometimes use Grok Imagine directly, and it gets the prompts and any images we give it.
A note you write in Studio doesn't reach the agents making your video until Nathan Amick accepts it. Nathan's own assistant, Grok Bot, may read notes, video details and upload information as soon as they're written, so he can manage the work. You see only the cuts and messages Nathan sends, and not the agents' working notes.
If we publish to your accounts, or use another tool to share files with you, that platform's own policy applies to what it receives.
What each AI service receives:
- Anthropic (Claude) and OpenAI, for AI agents: the transcript, plan and script, your notes, titles and descriptions, and sometimes stills or contact sheets from the video. We may use either on a video. Neither gets the audio or the full video.
- OpenAI, for thumbnails: text prompts, which can describe your video.
- ElevenLabs: the audio of your recordings and words to listen for, such as names. If we make sound effects, it gets the text prompts.
- fal.ai: text prompts, and sometimes a still image as a reference.
- Grok Bot: client notes and the transcript words they're pinned to, client, channel and video names, titles and descriptions, version and upload details, and sometimes the name of who wrote a note. It doesn't get email addresses, billing details or video files.
Each service's own terms govern how it handles that data, including whether it keeps the data or trains on it, and they may change. We make no promise on either point. Read them at Anthropic, OpenAI, ElevenLabs, fal.ai, and Grok Bot with Cursor. Clerk, Convex, Stripe, Vercel, Cloudflare, Google, Meta and GitHub post their policies on their own sites.
YouTube
Connecting a YouTube channel is optional. If you want us to publish your finished videos to your own channel, a member of your client account connects it in the channel's YouTube tab on Studio's Channels page. Reviewers can't connect or disconnect a channel.
Google asks you to approve three permissions, and Studio connects only when all three are granted:
- youtube.upload lets us upload your finished video to your channel, with its title and description, and set its thumbnail. Uploads start private.
- youtube.readonly lets Studio read your channel's id, name and handle when you connect. Studio shows them so you know which channel is connected. Before each upload, our production system reads them again to check it's publishing to that channel. After an upload, it reads that video's privacy status to check it's private. When a member of your client account opens a video's page on Studio, Studio reads the privacy status of the videos we published for it, at most every few minutes, so the page shows whether each one is private, unlisted or public.
- youtube.force-ssl lets Studio upload captions to the videos we publish for you. It lists a video's caption tracks, then adds ours or replaces the copy it added before. Our production system uses it for one other thing: if YouTube reports a new upload as anything but private, it sets that video back to private. Google describes this permission broadly, as seeing, editing and deleting your videos, ratings, comments and captions. We use it only on the videos we publish for you, and only for those two things.
Studio stores your channel's id, name and handle, the permissions you granted, any upload defaults you set, such as language and category, and the long-lived token Google issues. It encrypts that token before storing it. When we publish, our production system on Amick Labs computers asks Studio for an access token, which lasts an hour at most, and uses it for the upload and the checks above. Studio doesn't store those access tokens. Studio uploads captions itself.
We use this data only to publish your finished videos, thumbnails, titles, descriptions, chapters and captions to your own channel. We don't sell it, use it for ads or use it to train AI models. Studio doesn't send your Google token or channel details to the AI services listed above. Apart from Convex, which hosts Studio's database, we share it with no one except Google when we publish, or when the law requires it.
You can disconnect at any time in the channel's YouTube tab on Studio's Channels page. Studio then deletes the stored token and asks Google to revoke it. You can also remove Studio's access yourself at myaccount.google.com/permissions. After you disconnect, Studio keeps a record of which channel was connected until you ask us to delete it. Videos already on your channel stay there.
Studio's use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements.
Studio uses YouTube API Services. By connecting a channel, you agree to the YouTube Terms of Service. The Google Privacy Policy covers what Google does with your data.
Connecting an Instagram account is optional. If you want us to post your finished videos to Instagram as Reels, a member of your client account connects it in the channel's Instagram tab on Studio's Channels page. Reviewers can't connect or disconnect an account. It has to be a professional account, Business or Creator. Studio won't connect a personal account. You don't need a Facebook Page.
Meta asks you to approve two permissions, and Studio connects only when both are granted:
- instagram_business_basic lets Studio read your account's id, username and account type when you connect. Studio shows the username and type so you know which account is connected, and checks the type to turn away personal accounts.
- instagram_business_content_publish lets Studio post your finished videos as Reels. For each one, Studio gives Meta a signed link to the video file, which works for two hours, and asks Meta to make a media container from it, with the caption and cover we chose. Studio checks the container about once a minute until Meta has processed the video, then publishes it and reads the new post's link on Instagram.
Unlike YouTube, Studio posts to Instagram itself. Our production system on Amick Labs computers uploads the finished cut to Studio and asks Studio to post it, but it never receives your Instagram token. It does get your account's id and username, so it knows where a video is going.
Studio stores your account's id, username and account type, the permissions you granted, and the long-lived token Meta issues. It encrypts that token before storing it. Meta's tokens expire, so a daily job refreshes any token that will expire within 30 days. If Meta refuses the refresh, or the token runs out first, Studio deletes the token and you'll need to connect again. For each Reel it posts, Studio keeps the caption and cover settings, Meta's id for the post, its link, and whether it worked.
We use this data only to post your finished videos to your own account. We don't sell it, use it for ads or use it to train AI models. Studio sends your Instagram token to no one but Meta. Apart from Convex, which hosts Studio's database, and our production system as described above, we share your account details with no one except Meta when we post, or when the law requires it.
You can disconnect at any time in the channel's Instagram tab on Studio's Channels page. Studio then deletes the stored token. Meta offers no way for Studio to revoke a token, so to remove Studio's access on Instagram's side, go to instagram.com/accounts/manage_access. After you disconnect, Studio keeps a record of which account was connected, and of the Reels it posted, until you ask us to delete it. Reels already on your account stay there.
The Instagram Privacy Policy covers what Meta does with your data.
Who can see it
- Members of your client account can see that client's videos, uploads, notes and billing.
- Reviewers on your client account can see the client's name, all its videos and transcripts, and all its notes, and can leave or answer notes. They can't see billing or uploads.
- Your own agents, through keys you create, can do most of what a member can, but not billing.
- If you connect your own Grok Bot, Studio sends it each new version's title and a link to its page.
- Nathan Amick can see all of it.
- Files play through signed links that expire: six hours for cuts and one hour for previews of raw clips. Anyone holding a copy of a link can use it until then.
We share information beyond the services above only if you ask or the law requires it.
Cookies
amicklabs.com sets no cookies.
Studio uses Clerk's cookies to keep you signed in. Your browser also stores which client you last selected. Studio has no analytics or advertising cookies. Stripe's own pages may use their own cookies when you pay or manage billing. A video's page shows your published YouTube video in YouTube's privacy-enhanced player (youtube-nocookie.com). YouTube may store information in your browser when it plays, under Google's policy.
How long we keep it
We keep your account, videos and footage until you ask us to delete them. An upload you don't finish is removed after 7 days without activity. We keep billing records as long as tax and accounting rules require. Stripe keeps its own records under its own policy.
Your choices
Email us to see, correct or delete your information, and we'll answer within 30 days. You can disconnect YouTube or Instagram yourself, as described under YouTube and Instagram. Studio has no delete button, so we do deletions by hand, across Convex, R2, Clerk and the toolkit's computers. Deleting doesn't reach videos already published to your own accounts, or records we have to keep.
Studio is for adults, and we don't knowingly collect information from children under 13.
Changes
If we change this policy, we'll update the date above and email current clients before a change that matters takes effect.