Privacy policy
This policy explains how Artide Innova SRL handles personal data when you use rushbin: the team website run by Artide (today rushbin.artide.io, later rushbin.io) and the rushbin apps for Mac and Windows distributed by Artide. It is written under the EU General Data Protection Regulation (GDPR, Regulation 2016/679) and the Italian Privacy Code (Legislative Decree 196/2003).
1. Who we are
Artide Innova SRL, Via Rosina Sessa 16, 84128 Salerno, Italy, VAT number IT06198200658, REA SA-504766. Email hello@artide.studio, certified email (PEC) artideinnova@pec.it. We have not appointed a Data Protection Officer: the GDPR does not require one for our processing. Write to the addresses above for anything about your data.
2. Two different roles
Artide is the controller of the data needed to run the service itself: your account, the security of the site, the website statistics you consent to, and our relationship with the companies that buy rushbin.
Artide is a processor for the content that a team puts into rushbin: the catalog of its drives, previews, light copies, transcripts and everything extracted from its footage. For that content the controller is the company that owns the team (our customer), and we process it only on its instructions, under the data processing agreement. If you appear or speak in footage cataloged with rushbin, the company that filmed you or keeps the footage is the one responsible towards you: see section 7.
3. What we process as controller, why, and for how long
| Data | Purpose | Legal basis | Kept |
|---|---|---|---|
| Account: email, name, password (stored only as a bcrypt hash), role and teams | Giving you access to your team | For the members a customer invites, Artide's legitimate interest in providing the service the customer bought (art. 6.1.f); contract (art. 6.1.b) only when you are yourself the customer | Until the account is removed from all its teams, or the customer's contract ends (section 6) |
| Sessions: a random token (stored hashed), the name of the computer for the apps, when it was last used | Keeping you signed in | As for the account | 30 days for a browser session; an app session until you sign out or are removed, or until it has not been used for 90 days: then it expires and is deleted |
| Invite and password reset links and the codes that connect a Mac to the team (stored hashed), with the invited email and role, or the person and team a code is for | Joining a team, setting a new password, connecting the app without typing a password | As for the account | Each works once. Invites and the password links an owner makes: 7 days; password links sent by email: 1 hour; codes that connect a Mac: 10 minutes. Then deleted |
| Sign-up to create a team, only where the site offers it: the team's name, your name, email and password (stored only as a bcrypt hash), and the link that confirms the email (stored hashed) | Creating the team once you confirm that the email is yours | Steps you ask for before a contract (art. 6.1.b) | Until you confirm it, then it becomes your account and your team; a sign-up not confirmed is deleted after 7 days, or earlier when the same email signs up again |
| The date the owner who created a team by sign-up accepted the terms of service and the data processing agreement | Showing that the terms were accepted | Contract (art. 6.1.b) and legitimate interest in proving it (art. 6.1.f) | As long as the team exists |
| Emails the site sends: invites, password links and sign-up confirmations, with the recipient's address, the team's name, the name of the member who invites or makes the link, and the link | Bringing the link to the person it is for | As for the account, or as for the sign-up | The site keeps no copy of the emails; they go through Amazon SES (section 5) |
| IP address and email of failed sign-ins; IP address of demo visits, of tidy-ups with AI, of sign-ups and of requests for a password link; the addresses the site emailed | Blocking password guessing and repeated requests | Legitimate interest in security (art. 6.1.f) | Only in memory, 15 minutes |
| Error logs of the server, which may contain file names or email addresses | Running and fixing the service | Legitimate interest (art. 6.1.f) | Rotated by size: at most 50 MB for each of the site's services, the oldest lines overwritten first. There is no time limit: how long a line stays depends on how much the service writes |
| Website statistics through Google Analytics, loaded by Google Tag Manager, only if you accept them in the cookie banner | Understanding how the site is used, to improve it | Consent (art. 6.1.a and art. 122 of the Privacy Code) | See the cookie policy |
| Customer contact and billing data | Contracts, invoices, tax duties | Contract and legal obligation (art. 6.1.b, 6.1.c) | 10 years, as Italian law requires for accounting records |
Your email and your role are entered by the owner of the team who invites you; you choose your name and password yourself when you join. If you create a team by sign-up, you give all of them yourself. The rest comes from your use of the site.
Giving your account data is necessary to use the site; without it you cannot sign in. We do not sell personal data, we do not use it for advertising profiles, and we do not train AI models on it.
4. The apps
The rushbin apps work on your computer. They read your drives and keep the catalog on the computer; drives are only read, except for the moves you confirm on that same drive (duplicates into a trash folder, or a reorganization you accept). Search by content (speech, on-screen text, what is in the picture) runs entirely on your computer, and only for the drives where you turn it on.
If you use Tidy up with AI and confirm it for a drive, the names of that drive's folders and files with their sizes, dates, durations and codecs, never the contents of the files or frames, go to an AI service to receive a proposed reorganization. They go one of two ways, and the app names the way and who receives them before you confirm:
- When the app is connected to a team whose site offers it, the app sends this description to that site, which passes it to the site's own AI with the site's own key and returns the proposal to the app. The site keeps nothing of the description and does not write it in its logs; it keeps only how many tidy-ups each team ran in the current month, deleted when the month ends. On Artide's site that AI is always Anthropic (Claude): this is customer content, which Anthropic handles as our sub-processor (see the data processing agreement). Under Anthropic's commercial terms, Anthropic may not train models on it. Under Anthropic's rules for the API, it deletes the description and its answer within 30 days, unless the law requires it to keep them longer or its safety systems flag them as a breach of its usage policy: then it may keep them up to 2 years, and the safety classification scores up to 7 years. On a site run by someone else, whoever runs it chooses the AI and answers for it.
- Otherwise the app sends it itself to the AI you chose under Advanced, with your own API key: Anthropic (Claude), OpenAI, Google (Gemini), or a compatible service whose address you enter (for example Mistral, or a model running on your own computer). You are then the one who sends this data to that provider, under your agreement with it, and its terms decide how it keeps and uses the data; Artide does not receive it. Your key stays only on your computer, goes only to that provider and never reaches the team site.
Apart from Tidy up with AI, data leaves your computer only in these cases. First, when you connect the app to a team: then it sends that team's site the catalog, the previews, what was extracted for search, and the light copies for the drives where you turned them on. Each drive has a switch, "Send to the team": when it is off, nothing of that drive leaves the computer, Tidy up with AI works for it only with your own AI key and never through the team site, and the site deletes at once that drive's catalog, previews, analyses, transcripts, search vectors, light copies and original files, keeping only the data of clips that are also on other drives of the team. Decisions about clips that are only on drives kept on the computer stay on the computer. Second, the first time you use search by content, the app downloads the speech models from Hugging Face (huggingface.co); on Windows also the image model, and on a Mac in no team also the text model of SigLIP 2, which turns the words of a search into something comparable with the pictures. Like any download, Hugging Face sees your IP address. No other data is sent to it. Third, for automatic updates: an app connected to a team asks that team's site, every 6 hours, whether a new version of rushbin is available; an app in no team asks this of Artide's site (rushbin.artide.io) only if you allow it when the app asks, and you can stop it in Team. On the Mac the updater (Sparkle) sends with each request the version of rushbin and of the updater and the languages set on the Mac, as browsers do; on Windows the request carries neither. Nothing about you or your files is sent, and our site does not keep the internet address it answers to. When there is a new version, its installer is downloaded from the release storage of the site that offered it, through a link that works for at most an hour: on the Mac the updater downloads it by itself and installs it when you quit rushbin; on Windows the app shows the new version and you download it. On Artide's site the release storage is Cloudflare R2, which, like any download, sees your IP address.
What is automatic. Transcripts are written by Whisper (large-v3-turbo); text on screen and labels are recognized by Apple Vision on the Mac and by Windows OCR on Windows; search by looks compares the pictures with the words of a search through SigLIP 2. All of this runs on your computer, except that on a team site the words of a search are turned into numbers on the site's server. These are AI models and they make mistakes: Whisper can write sentences nobody said, especially over music, crowds and choirs, or repeat a line. The app and the site say so where they show the results ("Automatic transcript (Whisper). May contain errors.", "Recognized automatically."), mark them as made by AI in the data they provide, the team export included, and let whoever can decide hide a wrong line of a transcript and mark a transcript as checked, with who and when. rushbin does not recognize faces or emotions, does not identify people, and does not put people into categories by age, sex or other physical features. Tidy up with AI only proposes: you choose each move, every move can be undone, and nothing is deleted. rushbin has no chatbot.
5. Who else handles the data
| Provider | What for | Where the data is | Transfers outside the EU |
|---|---|---|---|
| Amazon Web Services EMEA SARL, Luxembourg (AWS Lightsail and Amazon SES) | The server that runs the site and its database (Lightsail); sending the site's emails: invites, password links and sign-up confirmations, with the recipient's address, the team's name, the name of who invites or makes the link, and the link (SES) | Frankfurt, Germany (eu-central-1) | Access by Amazon group companies in the United States: EU-U.S. Data Privacy Framework and standard contractual clauses in AWS's data processing addendum |
| Cloudflare, Inc., United States (R2 and DNS) | Storage of light copies, requested originals and database copies, and of the rushbin installers (R2); DNS for Artide's domains | R2: buckets with the EU jurisdiction (data stored in the EU) | Cloudflare, Inc. is in the United States: EU-U.S. Data Privacy Framework, with the 2021 standard contractual clauses as fallback, in Cloudflare's data processing addendum |
| Google Ireland Ltd (Google Tag Manager and Google Analytics), only with your consent | Website statistics | EU and United States | EU-U.S. Data Privacy Framework (Google LLC) |
| Anthropic Ireland, Limited (Claude API), only when a member uses Tidy up with AI on Artide's site | Proposing a reorganization of a drive from the names of its folders and files (section 4) | Outside the EU | Standard contractual clauses in Anthropic's data processing addendum |
Each provider acts under a data processing agreement. The EU-U.S. Data Privacy Framework is the one the European Commission found adequate (Decision 2023/1795); the standard contractual clauses are those of Commission Decision 2021/914. You can get a copy of the standard contractual clauses that cover a transfer by writing to us at the addresses in section 1. The current list of providers that handle customer content is in the data processing agreement. An AI you choose for Tidy up with AI with your own key is not on these lists: the app sends it the description directly, under your agreement with that provider (section 4).
6. When a team or an account ends
When a person is removed from a team, their sessions in that team end; the decisions they took keep their name, as part of the team's history. The owners of a team can download an export of the team's data from the site at any time: a ZIP with the catalogs, previews, decisions with who took them and when, on-screen text and labels, transcripts, search vectors (in JSON), requests for originals, members and their roles (no passwords or sessions) and the records of removals on request; light copies come as download links that work for 7 days (terms of service, clause 12). When a drive is removed from a team, the decisions and the records of removals on request stay; a decision may contain the path and the identifier of the removed drive. When a customer's contract ends, we delete the team's data within 30 days; if the customer switches provider or asks for an export, at the end of the period in which it can retrieve its data (terms of service, clause 12). Database copies roll over within a further 31 days.
7. If you appear in footage
rushbin catalogs footage that our customers shot or keep. If you are filmed or recorded in that footage, the customer decides why and how it is kept, and it is responsible for informing you and for the legal basis (for example a release you signed, or its legitimate interest in keeping an archive of its own productions). Please contact that company first. If you do not know who it is, or you need help, write to us and we will pass your request to the customer, as the data processing agreement requires. rushbin does not recognize faces or identify people.
When the owner of a team removes a clip's data at the request of a person in it, the site at once deletes its previews, what was seen and said in it, its search vectors, its light copy and any original; the clip's name stays in the catalog, marked as removed. What was in the database stays in the nightly database copies until they expire, at most 31 days later.
Transcripts, on-screen text and labels are made automatically by AI models and may contain mistakes (section 4). The members of the team who can decide can hide a wrong line of a transcript; a hidden line is no longer shown to the others or found by search.
8. Your rights
You can ask us for access to your data, its correction or deletion, the restriction of processing, data portability, and you can object to processing based on legitimate interest (articles 15 to 22 GDPR). Where processing is based on consent, you can withdraw it at any time, without affecting what happened before: for cookies, from the cookie settings link at the bottom of the site. We make no decision based solely on automated processing that produces legal effects concerning you or similarly significantly affects you (art. 22 GDPR). We answer within one month. You can also lodge a complaint with the Italian supervisory authority, the Garante per la protezione dei dati personali (garanteprivacy.it), or with the authority of the country where you live or work.
9. Security
Connections are encrypted (HTTPS). Passwords and tokens are stored only as hashes. Every team's data is kept apart by every query and by row level security in the database. Links to videos are signed and expire within an hour. More details are in annex 2 of the data processing agreement.
10. Changes
We will publish any change here with a new version number and date; for important changes we will tell the owners of the teams in advance.