Getting things ready…
Back to homeYour ideas. Your choice.
Confetti privacy notice · Last updated: October 22, 2026
What we collect
When you sign in with Google, we receive your account identifier, name and email. We store your chosen username, maker profile, and the projects and conversations you save. Your username and public profile are visible to others; your email, private projects and conversations are not published.
How your projects are processed
We use Supabase for accounts, our database and private file storage, and Vercel to host Confetti. Your projects, saved versions, part notes and the files you attach are stored privately with your account. Part notes are kept until you delete the note, the project or your account; when you empty a note we keep only which part it was on. Notes that an earlier version kept only in this browser move to your account when you open that version. To generate or edit a design, we send the current prompt, full saved conversation, design, relevant project files, maker preferences, attached files and any photos to third-party AI service providers that generate text, code, images and 3D models for us. Inspiration images and concept pictures are made the same way. If a provider is unavailable, we may send the same request to another provider of the same kind. These providers process requests to return results to Confetti, under their own terms and retention practices; some keep request data for a limited time, for example for abuse monitoring or so we can recover results after an interruption. You can ask support which service providers currently process your data.
Pending requests, including attached photos, are saved only on this device until accepted. A request we refuse stays there until you edit or discard it. If any are left when you sign out, we list them and delete them only if you choose to. Background generation stores the input and intermediate stages privately on our server so it can recover from interruptions. Photo copies in the request payload and intermediate stages are removed when the request finishes, fails or is canceled, with one exception: when a request fails or needed automatic fixes, we keep its intermediate stages privately for up to seven days to find and fix the problem, then delete them, unless research data sharing is on for your account (see below). While a design request runs, its plan and newest unchecked draft are saved with the request so you can watch it; if it fails, the last draft stays so you can view it. Both are deleted with the request’s other details after seven days. Abandoned requests expire after 24 hours. Delivered request payloads are cleaned up after seven days; undelivered results are retained so you can recover them. Images sent with design requests and other files you attach to a message, such as 3D models and PDFs, are stored privately with that project so Confetti’s design assistant can inspect earlier references across sessions. Saved images are the resized copies sent by your browser with camera metadata removed. The concept image you select for a project is kept with that project on your account, and Confetti adds it to the project’s later design requests from any device until you continue without an image. These references are not part of model versions, downloads or public shares, and are deleted when you delete the project or your account. Do not include secrets or information you do not have permission to submit. Technical service logs may include network and request metadata used to operate and secure the service.
We make an encrypted backup of our database and stored files every day and keep it with a separate cloud storage provider. We use backups to restore the service after a failure. If you delete your account and then ask support within 90 days, we may also use them to recover your projects for you; Confetti has no other way to restore a deleted account. While daily backups run normally, each one is deleted within 95 days. If backups stop, we keep the last three until new ones replace them.
Credits and payments
We store your credit balance, subscription status, and credit usage to operate your plan. Billing records contain operation types and token counts, not copies of your designs or prompts. When paid checkout is enabled and you use it, Stripe receives your account email and payment information to process subscriptions, print purchases or electronics purchases. Card details are handled by Stripe; they are not stored in Confetti’s project database. Referral links use a signed cookie for up to 30 days to remember your invitation. We record referral attribution and credit rewards privately; referrers see aggregate counts, not your email or payment details. A private fingerprint of your email remains after deletion to prevent repeated welcome-credit and referral claims, and for about two months to prevent a second grant of the same month’s free credits. Financial records may remain with the payment provider after account closure. We also retain print and electronics order IDs, amounts and payment/fulfillment activity separately from projects for payment reconciliation and order support; the account link is removed on account deletion.
Operational diagnostics
When connected, Sentry receives limited error diagnostics to help us identify and fix failures: the affected feature, error category, code location, release, and random diagnostic or job/order reference. We remove exception messages and request contents before sending. We do not attach prompts, photos, designs, email addresses, cookies, authorization headers, or session recordings. Browser requests to the diagnostic service necessarily expose network metadata to that service; we do not intentionally include an IP address in error reports, and the service is set not to store IP addresses. These service diagnostics operate separately from optional product analytics. If you contact support, you can include or remove the displayed diagnostic reference so we can connect your report to the failure.
We summarize account creation, successful sign-ins, generation outcomes, credit usage, and order records in a private operational dashboard. Sign-in records are retained for 90 days. This uses records needed to operate the service and is separate from optional browsing analytics.
Optional analytics
When configured, PostHog receives limited product-usage events to help us improve Confetti. Analytics are off until you allow them, wherever you are; nothing is sent to PostHog before you choose Allow analytics. A choice to turn analytics off is always respected. We measure actions such as opening the workbench, completing generation, and exporting a project, using a browser-generated identifier and, after sign-in, an opaque account identifier so we can measure repeat use across devices. We also classify broad referral sources and distinguish internal testing from customer activity. We time how long the 3D preview takes to load, with file sizes, part counts and a coarse device class (rounded CPU cores and memory, and the browser's network type), but never the model itself. Confetti also keeps these preview timings for 90 days, without any account or network identifier. When you open the preview of your own design, the timing is linked to that design run, so our team can see how fast each run's preview loaded. When a signed-in browser is first linked to your account identifier, PostHog also stores, once, where that visit began: the full address of the page you opened (which can include a project or design identifier and link details such as an invitation code), the referring website and page, any search terms the referring search engine passed along, and campaign tags in the link. We do not send prompts, photos, design contents, email addresses or usernames to analytics. PostHog also receives network metadata when your browser connects to it; we disable geolocation enrichment. We use analytics to improve our own service, not for advertising or cross-site tracking. Automatic click and text event collection is disabled. You can change your choice below at any time; we also honor browser Do Not Track and Global Privacy Control signals. Essential account/session storage still operates when analytics is off.
If you allow analytics, PostHog also records sessions: the page layout, scrolling, clicks and navigation, so we can see where Confetti is confusing or broken. Recordings hide all text on the page and everything you type, and do not include your photos, generated images, the 3D preview, browser console or network contents. Page addresses keep only the section name, never project identifiers, usernames or query details. Turning analytics off stops recording. Session recording began on October 18, 2026; if you allowed analytics before then, that choice now also covers recording, and you can change it below.
Sharing and external services
Publishing is an explicit action. It exposes the selected design, build content, caption, creator profile and remix attribution. It does not publish your chat, photos, other versions or model-provider traces. Downloads and copies already made by others cannot be recalled. Retailer and fabrication services have their own privacy policies. A printing file is sent to a print service only when you request a quote, estimate or order from it. Slant 3D receives the printable STL files and selected print options when you request its live quote. When you calculate shipping, Slant and its shipping providers receive the recipient name, address and email you enter. You pay Confetti by Asteria for that order through Stripe, which processes the payment for us. We do not store your delivery address in your project or send it to analytics. The checkout expires after one hour; provider retention follows their own policies.
When you ask for print estimates and agree to share the file, our manufacturing partner, Craftcloud, receives the selected STL file and the destination country. Where Confetti offers checkout for those prints, Craftcloud and its manufacturer also receive the recipient name, phone number and U.S. delivery address you agree to share; Confetti places that order from its own Craftcloud account and you pay Confetti through Stripe. Files you upload on Craftcloud’s own site go directly to Craftcloud. Links to parts retailers open their websites; we do not send them your details. Product thumbnails load directly from supported retailer image hosts with no referrer. Those hosts receive your IP address and the image request, but not your project text, files or account details.
When you request an electronics shipping quote, Mouser receives your selected component numbers, recipient name, U.S. delivery address and phone number. We keep delivery details privately while preparing the order and remove them from the active order record after supplier acceptance, sandbox payment completion, checkout expiry or account deletion. Abandoned unpaid quotes are scrubbed within 24 hours of expiry. Orders that need support review retain the details needed to resolve them. Mouser and Stripe maintain their own records under their respective policies. Order history in Confetti includes selected part numbers, totals, payment status and shipment tracking, and is not public.
Resend processes transactional emails, including account verification, a signup welcome, and order updates. Support alerts send the ticket category, reference and first 1,000 characters of your submitted message to our support team. Order messages include your email address, order reference, purchased project title, payment or refund totals, and available supplier tracking details. Email content is retained in our delivery queue for up to 30 days after processing; minimal delivery and duplicate-prevention records remain with the order or account. Account phone numbers are not requested; delivery phone numbers are collected at physical-order checkout.
Support and access
Support requests contain the message you submit, up to three photos you choose to attach (for example, of a damaged or wrong order) and any public design you report. Authorized support staff can read and reply to them and view the photos. Photos are stored privately and are deleted with the request or your account. Private designs are not automatically attached.
A small number of authorized Confetti staff can also open private projects, including the prompts and messages you send, Confetti's replies and the project files, to operate, secure and support the service: for example, to find out why a design request failed or was slow, to look into abuse, or to answer your support request. They do not use this access for any other purpose, except the product research described next. Except through research data sharing, described below, we do not sell your private project data.
From October 21, 2026, Confetti also uses the prompts and conversations you send on or after that date to learn what makers build with Confetti and to improve it, for example to count which kinds of projects are most common. Authorized staff may read them for this purpose, and the same AI service providers described above may sort them into categories for us. We keep the results inside Confetti, and any results we show outside Confetti are totals that contain no prompts, names or other identifiers. Conversations sent before October 21, 2026 are not used for this research.
When Confetti refuses a request under the community rules, for example a request for sexual content or a weapon, we keep the refused message, its reason and your account for 90 days so our moderation team can review it. We do not keep photos or the rest of the conversation with it. The record is deleted with your account.
Research data sharing
Research data sharing is part of Confetti’s Free plan, and it is on by default for every account created in the United States on or after October 10, 2026. If you are 16 or older, Confetti may license your saved prompts, conversations with Confetti and designs, including generated design files and specifications, to companies that build or study AI models, to train and evaluate those models, in exchange for payment. Paid plans can turn it off. Under California and other US state privacy laws, this is a sale or sharing of personal information.
The licensed data also includes: for requests that fail or need automatic fixes, the request and Confetti’s intermediate design attempts with the reasons they were rejected; the inspiration images generated for your requests, the prompts used to make them and which AI provider made each, which image you chose or skipped (or the idea you wrote instead) and how long you took; and, for designs, which design engine built each version and what you did with it afterwards, such as downloads, version restores, build reports, print and kit orders and publishing.
Before any licensing, we remove your name, email, username, account and project identifiers, photos you upload, delivery addresses, phone numbers and payment details. Support messages and order details beyond what you ordered are not included. Prompts and conversations are free text, so they can still contain details you typed: leave out personal information you do not want shared. We do not share data from people who tell us they are under 16, from accounts created before October 10, 2026, or from people outside the United States. If you agreed to research sharing before October 19, 2026, when it covered the Free plan only, nothing you create while on a paid plan is included unless you turn it on in Settings. We will require recipients not to try to identify anyone from the data.
Do Not Sell or Share My Personal Information
On a paid plan, you can turn research sharing off anytime in Settings while signed in. It stays off while your paid plan lasts; if your account returns to the Free plan, sharing resumes for data you create after that. If you live in California, or in another state whose law gives you the right to opt out of the sale or sharing of your personal information, you can opt out on any plan: contact support or email support@confetti.build and ask us not to sell or share your personal information. We will turn research sharing off for your account on every plan. We treat a browser’s Global Privacy Control signal as a request to turn it off. If you turned research sharing off before October 19, 2026, it stays off. Turning it off stops your data being included in future licensing and deletes the records we kept for it; data licensed before then cannot be recalled. Deleting a project or your account removes it from future licensing. In the past 12 months, Confetti has not sold or shared personal information.
Children
Confetti is not meant for children under 13, and you must be at least 13 to create an account. If you are under 18, use Confetti with a parent’s or guardian’s permission. We never include data from people who tell us they are under 16 in research data sharing. If you believe a child under 13 has given us personal information, contact support and we will delete it.
Keeping and deleting information
Saved account data stays until you delete it or your account. Delete your account in workspace settings to remove its cloud projects, profile, publications, interactions and support requests. The current browser’s account cache is cleared. Other devices may retain local copies until you clear their site data. Deletion from the active database does not instantly erase provider logs, routine backups or copies other people downloaded. Deleting your account also removes your profile images and the files of your community uploads that no one else’s remix or publication still uses. Deleted data leaves our encrypted backups as described above. For 90 days after you delete your account, we keep a private record of it: its identifiers, email, creation and deletion dates, number of design requests, and the name, identifier and dates of each project. It does not contain designs, files, chats or prompts. We keep it because we have a legitimate interest in answering order questions, payment disputes and abuse reports about the account, and we use it only for those. Our backups contain this record too, so it leaves them within about 185 days. To have it deleted sooner, contact support; we will delete it unless an open order, dispute or abuse report needs it. Provider logs are subject to their separate retention policies.
You can download each project as a ZIP file. For a copy of your other account data, access, corrections, privacy questions or deletion problems, contact support or email support@confetti.build.
Who we are and changes
Confetti is provided by Asteria Labs, Inc. (“Asteria”), 2261 Market Street STE 74309, San Francisco, CA 94114, which is responsible for the personal information described here. If we make a material change to this notice, we will update the date above and, where appropriate, tell you in Confetti or by email before it takes effect.