Privacy Policy

1. Who we are

Kondense is a service operated by Simon Chawla, carrying on business as Littlefoot Digital ("Kondense", "we", "us"). We are the data controller for the personal information described in this policy. This policy covers kondense.app and the application at studio.kondense.app.

You can reach us about anything in this policy at support@kondense.app, or by post at 242 Bryanston, Lucasville NS B4B 0N2, Canada.

For the account you hold with us — your name, email, billing details and usage — we decide how the information is handled, so we are the controller. For the video content you bring to the Service, we act on your instructions: you decide what to upload and what to do with it. If you are a business that needs a data processing agreement covering that content, ask us and we will provide one.

2. What we collect

Information you give us

• Account details — your name, email address, and a password, which we store only as a salted hash and never in readable form.
• Social sign-in identifiers — if you sign in with Google or Facebook, the identifier they give us for you, plus the name and email address on that account.
• Billing details — your Stripe customer and subscription identifiers, your plan, and the card brand and last four digits. We never receive or store your full card number, expiry date or security code; card entry happens on Stripe’s own hosted page.
• Cancellation feedback — if you cancel a plan and tell us why, we keep the reason and anything you write.
• Correspondence — emails you send us, and your conversations with us in the support chat, with anything you attach to them.

The content you bring

• Source video — the file you upload, or the video we retrieve from a link you give us, together with the audio track we extract from it.
• Transcripts — the full text of everything spoken in your video, with per-word timings, detected language, speaker labels and sentiment scoring.
• Project material — project names, chapter and clip titles, descriptions, captions, caption styling, images and overlays you add, and the finished clips we render.
• Media library — file names, sizes and types of images and videos you keep in the Service.
• Derived analysis — points of interest, detected shot boundaries, framing decisions and subject masks that the Service computes from your video in order to crop and follow the action.

Connected social accounts

If you connect YouTube, TikTok, X, LinkedIn or Instagram so we can publish for you, we store the account identifier, display name, handle and avatar image URL from that platform, together with the access and refresh tokens that let us post, their expiry, and the permissions you granted. Disconnecting an account removes those tokens.

Information we collect automatically

• IP address — recorded when you register and kept against your account, to protect the Service and the free trial from fraud and abuse, and, if an Instagram, Facebook or TikTok ad brought you, to report to that platform (see section 3). It is also stored with your login session.
• Ad click details — if you come to us from one of our ads on Google, Instagram, Facebook or TikTok, the click identifier that platform added to the ad’s link, the campaign tags on it, when you arrived and, for an Instagram, Facebook or TikTok ad, your browser’s user-agent string as you signed up, kept against your account. Instagram and Facebook add their identifier to links people tap outside ads too, so you may arrive with one without having seen an ad.
• Browser identifier — a random value kept in a cookie on your browser (see section 10) and against your account, to protect the Service and the free trial from fraud and abuse.
• Device fingerprint — a one-way value derived from technical characteristics of the device you use, kept against your account, to protect the Service and the free trial from fraud and abuse.
• Language — the language your browser asks for and the one you choose, kept against your account, so that we can show you the Service and write to you in your language, and so that we can see which languages people want next.
• Session data — a session identifier, your IP address and your browser’s user-agent string, for as long as you stay signed in.
• Operational logs and error reports — records of processing steps, failures and diagnostics, which include project, clip and video identifiers and may include project names.
• Support chat — our support chat is our own. An AI assistant answers first, from our help content, and someone from our team can take over at any time; you can look back on your earlier conversations in the chat. In the Service, nothing of it loads until you open it, with the chat button or, on a phone, Chat with us, or while you have a conversation going, so that you can see our replies, and never in the clip editor. On our website it loads with each page, and after a few seconds may invite you to ask us something; nothing about you is recorded unless you write to us. It records the messages you send and our replies, the page you write from and the page that brought you there, your browser and operating system, the country your IP address is in, and a one-way fingerprint of your IP address, which cannot be turned back into it, to stop the chat being abused. If you leave your email address there, we use it to answer you. While you are signed in it also has your name, email address and account, with your plan, the minutes you have left and the date you signed up, so we know who we are helping. The assistant runs on OpenAI (see section 5), which receives your messages, the conversation so far and, while you are signed in, those account details.

3. Why we use it, and our lawful basis

If you are in the UK, the EEA or Switzerland, the UK GDPR and EU GDPR require us to have a lawful basis for each use. Ours are:

• To provide the Service — creating your account, processing your videos into clips, storing your projects, publishing to accounts you connect. Basis: performance of our contract with you.
• To bill you — taking payment, managing subscriptions and credits, handling refunds and disputes. Basis: performance of our contract, and compliance with tax and accounting law.
• To send service messages — verification and password-reset emails, notice that your clips are ready or that a project or scheduled post failed, and notice that a connected account needs reconnecting. Basis: performance of our contract.
• To check in with you — a few short emails from our founder: one if you have signed up but not made any clips yet, one if you have made clips but not exported any, and one asking for your feedback once you have exported. We pick which to send from your account’s activity, send each one at most once, and some are sent automatically. Replies go straight to our founder. Every one has a link to stop them, which leaves the service messages above unaffected. Basis: our legitimate interest in helping you get value from the Service and in improving it.
• To help you when you ask — answering you in the support chat or by email, with your account in view so we can see what you see. Basis: performance of our contract with you, and, before you have an account, our legitimate interest in answering the people who contact us.
• To keep the Service secure and working — diagnosing errors, investigating abuse, and preventing fraud, including misuse of the free trial. Basis: our legitimate interest in a service that works and is not abused.
• To see which of our ads work — when one of our ads on Google, Instagram, Facebook or TikTok brings you to us, we tell that platform when you sign up, when your first clips are ready and when you make your first payment, with the click identifier its ad added, the time, and what the payment was. Meta (for Instagram and Facebook) and TikTok are also sent the IP address you signed up from and your browser’s user-agent string, which they match the click by. We never send them your email address, your name or your content. The platforms use these reports to measure our ads and to choose whom to show them to, under their own policies. Basis: our legitimate interest in knowing which ads bring people who go on to use the Service. You can switch this off at any time under Settings → Privacy, and an account opened in a browser that sends a Global Privacy Control signal starts with it switched off. Switching it off stops further reports; it cannot recall ones already sent.
• To improve the Service, and to train our models — understanding which pipeline steps fail and what they cost to run, and using the content you bring, including video frames and transcripts, to develop, train, test and improve the models that crop, follow speakers, detect shots and choose clips. Basis: our legitimate interest in improving what we offer. Paid plans can opt out, and you can object — see section 4.
• To comply with the law and to establish or defend legal claims — including responding to copyright complaints. Basis: legal obligation, and our legitimate interest in defending ourselves.

Where we rely on legitimate interests, we have considered the effect on you, and you can object — see section 11.

4. Automated processing and AI

Turning a video into clips is automated. It is worth being specific about which parts of that happen on our own machines and which involve an outside company.

On our own servers

All of the visual analysis runs on our own containers, on our own infrastructure, using models whose weights we hold. Your video frames are not sent to any third party for this. That covers:

• saliency prediction — working out where in the frame the interesting part is;
• face detection and a speaking-activity score, used to decide who to frame and when to cut between speakers;
• layout and component detection, and subject segmentation;
• shot-boundary detection; and
• rendering the finished clips.

Face detection here finds that a face is present and where it is. It does not identify anyone, does not match faces against any database, and does not create a face template or any other biometric identifier. We do not perform facial recognition.

Sent to an outside company

• Audio, for transcription. We extract the audio track and give ElevenLabs a short-lived link to fetch it. They return a transcript with word timings and speaker labels. The video picture is never sent.
• Transcript text, for analysis. We send OpenAI the transcript text, the project name, and — if you use the "most discussed" mode — the text of public comments on the source YouTube video. This is how clips are chosen and how titles, descriptions and chapters are drafted. No video, audio, images or frames are sent to OpenAI.

Training on your content

The analysis models listed above are ours, and we keep improving them. We may use the content you bring — including video frames and transcripts — to develop, train, test and improve those models and the rest of the Service. That is how the cropping, speaker-following, shot detection and clip selection get better over time.

This happens inside our own infrastructure. We do not sell your content, license it to anyone else, or release it as a dataset, and we do not use it to train models for anyone other than Kondense.

If you are on a paid plan, you can opt out under Settings → Privacy. From that point we will not use your content for training. Free accounts contribute to training, and that is part of what keeps the free tier free.

Separately from the plan you are on, if you are in the UK, the EEA, or anywhere else whose law gives you a right to object to processing we base on legitimate interests, you can exercise that right whatever you pay — see section 11.

Automated decisions

Nothing we automate produces a legal or similarly significant effect on you. Clip selection, framing and generated text are editorial suggestions that you review and can change or discard.

5. Who we share it with

We share personal information only with the providers below, only for the part each plays, and only to the extent it needs. We do not sell it or rent it, and beyond the ad reports described in section 3, we do not give it to anyone for advertising or their own marketing.

• Google Cloud (United States) — hosting, database, file storage and job queues. Everything we store lives here.
• Cloudflare — sits in front of the application, terminates the secure connection and filters abusive traffic. It sees your IP address and request metadata.
• ElevenLabs — transcribes the audio from your videos.
• OpenAI — analyses transcript text to select clips and draft titles, descriptions and chapters, and writes our support chat assistant’s answers from what you write there.
• Stripe — takes payment and manages subscriptions. Stripe handles your card details directly and is its own controller for them.
• Google Ads, Meta and TikTok — only if one of their ads brought you to us, that platform receives the reports described in section 3: the click identifier its ad added, when you signed up, got your first clips and made your first payment, and what that payment was, and, for Meta and TikTok, the IP address you signed up from and your browser’s user-agent string. Each is its own controller for what it receives.
• Resend — delivers our transactional email. Receives your email address and the content of those messages.
• Google Workspace — hosts our own mailboxes. The check-in emails in section 3 are sent from our founder’s mailbox there, and anything you email us, replies included, arrives there. It receives your email address and the content of those messages.
• Sentry — collects error reports and application logs so we can fix faults. Configured not to send personal information; it does receive stack traces, log lines and project identifiers.
• Intercom (United States) — ran our support chat before we built our own, and keeps the conversations from that time until we delete them.
• Klipy — powers the GIF and sticker picker. Your browser talks to Klipy directly, so Klipy sees your search terms, IP address and browser, together with an anonymous token we derive for you. Your video, audio and transcript are never sent to Klipy.
• Social platforms you connect — YouTube, TikTok, X, LinkedIn and Instagram receive the clip you chose to publish and the caption, title, description and privacy setting you set. What happens after that is governed by their own policies.
• Font and asset services — Fonts Bunny, Google Fonts and jsDelivr serve typefaces and emoji images to your browser and to our renderer. They see the request, not your content.

We may also disclose information where the law requires it, to respond to a valid legal request, to enforce our Terms of Service or investigate abuse, or to protect the rights and safety of our users or the public.

If Kondense is transferred to a company, merged or acquired — including the incorporation of the Canadian company that will operate it — information may transfer with the business. We will tell you before your information becomes subject to a materially different policy.

Staff access

A small number of administrators can see account records and, where it is necessary to investigate a fault or a support request, sign in to an account to see what you see. Every such action is recorded in an audit log with who did it, to which account and why. Administrators acting this way cannot change your profile or your billing.

6. Where your information goes

We are based in Canada, and our infrastructure runs in Google Cloud’s us-central1 region in Iowa, United States. Our other providers are in the United States or process information there. So if you are in the UK, the EEA, Australia or elsewhere, your information will be transferred outside your country.

For transfers out of the UK and EEA we rely on the European Commission’s Standard Contractual Clauses and the UK Addendum, incorporated into our agreements with those providers, together with the technical measures in section 8. Canada holds an EU adequacy decision for commercial organisations.

One exception worth naming: the queue we use to pass jobs between our own services is a global Google Cloud service and may hold messages in other regions. Those messages contain only identifiers and internal links — never transcripts, video or audio.

7. How long we keep things

• Rendered exports — deleted 30 days after they are produced. Download and keep anything you need.
• Uploads that never finished — deleted 24 hours after they stall.
• Editor undo history — deleted 24 hours after you stop editing.
• Projects and clips — kept while your account is open, so you can come back to them. When you delete a project, its clips and their exports are deleted straight away, and a record of the project itself (its name, settings and speaker names, and when it was created and deleted) is kept for up to 30 days for auditing, then deleted. Everything is deleted within 30 days of deleting your account. On the free trial, a project is kept for 30 days from when it was made (or, for a project we made for you, from when you claimed it), then deleted with its clips, exports and video unless you have chosen a plan by then; we email you a week before.
• Videos you upload, with their audio and transcripts — kept while your account is open, and deleted within 30 days of deleting your account. On the free trial, they are deleted with the project made from them, at the end of its 30 days.
• Account and profile records — kept while your account is open.
• Billing and tax records — kept for as long as tax and accounting law requires, typically six to seven years, even after your account closes.
• Registration IP records — kept to prevent abuse of the free trial.
• Ad click details — kept against your account while it is open, and removed from it when the account is deleted. What an ad platform has received from us is kept under that platform’s own retention policy.
• Browser identifiers — kept against each account while it is open, and removed from it when the account is deleted.
• Device fingerprints — kept against each account while it is open, and removed from it when the account is deleted.
• Language — kept against each account while it is open, and removed from it when the account is deleted.
• Email fingerprints — a one-way fingerprint of every email address an account has used, kept after the account is deleted, to prevent abuse of the free trial. It cannot be turned back into the address.
• Error reports and operational logs — kept for up to 90 days.
• Support conversations — deleted 12 months after the last message in them, together with what our support chat records about you, and within 30 days of deleting your account. Those from before we ran our own support chat, kept by Intercom, are deleted with your account, or when we close our Intercom account.
• Administrator audit records — kept for as long as we need them for accountability.

When we delete a file from storage, our storage provider keeps a recoverable copy for 7 days before it is destroyed, and copies may persist in encrypted backups for up to 90 days. Intercom likewise keeps what we delete there recoverable for 7 days.

8. Deleting your account

You can delete your account yourself, from Settings → Delete account. You will be asked for your password to confirm. If you have an active subscription, it is cancelled as part of the process.

When you delete your account we remove, within 30 days: your profile and login details; your projects, clips and rendered exports; the videos you uploaded, with their audio and transcripts; your media library; your connected social accounts and the access tokens we hold for them; your support conversations, with what our support chat records about you; the details of the ad click you arrived from; and your style templates and saved settings. Encrypted backups are cycled out within a further 90 days.

We keep, after deletion: billing and tax records we are legally required to retain; registration IP records, which are not linked to your identity; a one-way fingerprint of your email address, kept to prevent abuse of the free trial, which cannot be turned back into your address; and records of copyright complaints or abuse investigations where we need them to defend a legal claim or prevent repeat abuse.

Deleting your account does not remove clips you already published to a social platform. Remove those on the platform itself. It also does not reverse improvements our models have already learned from your content — see section 4, which explains why that cannot be undone and how to stop it happening in the first place.

9. How we protect it

• Traffic to and from the Service is encrypted in transit with TLS, and data is encrypted at rest by our cloud provider.
• Passwords are stored as salted bcrypt hashes. We cannot read them, and we will never ask you for yours.
• Your media is stored in a private bucket with public access blocked at the provider level. Your browser reaches it only through short-lived signed links that expire.
• Uploads go straight from your browser to storage; the file itself never passes through our application servers.
• Administrative access is restricted, gated behind a separate permission, excluded from browser caching, and logged.
• Internal services authenticate to each other, and incoming webhooks are signature-verified.

No system is perfectly secure, and we cannot guarantee absolute security. If a breach affects your personal information and creates a real risk to you, we will notify you and the relevant regulator as the law requires.

10. Cookies

We use only our own cookies: the strictly necessary ones and, if you pick a language, one that remembers it. There are no advertising or analytics cookies on our site or the Service, and our support chat sets none. The ad click you arrive with is kept on our servers with your session until you sign up, not in a cookie of its own.

• Session cookie — keeps you signed in. It is HTTP-only, sent only over a secure connection in production, and lasts up to 14 days.
• XSRF-TOKEN — protects forms against cross-site request forgery.
• kondense_language — remembers the language you chose, if you choose one on the Service. It holds nothing else and lasts a year.
• kondense_browser — holds a random identifier that helps protect the Service and the free trial from fraud and abuse. It is HTTP-only, sent only over a secure connection in production, and lasts up to 400 days.

Your browser also stores your light/dark theme choice locally, which never leaves your device. Once you have written to us in the support chat without being signed in, it also keeps a key to your conversations with us, which it sends only to our support chat, so that you can see our replies and look back on earlier conversations. The key is removed after 12 months without a message from you. If you close the chat’s invitation on our website, or open the chat, your browser remembers that for the rest of that visit only, so the invitation does not come back.

Blocking our own cookies will stop you being able to sign in. Third-party platforms you connect, and the font and GIF services named in section 5, may set their own cookies under their own policies.

11. Your rights

Wherever you live, you can ask us to:

• give you access to the personal information we hold about you;
• correct anything inaccurate or incomplete;
• delete your information — see section 8;
• give you a copy in a portable, machine-readable format, or send it to another provider;
• restrict or object to a particular use, including any use we base on legitimate interests; and
• withdraw consent where we relied on it, without affecting what we did beforehand.

To exercise any of these, email support@kondense.app from the address on your account. We will respond within 30 days, and will tell you if we need longer. We do not charge for this, and we will not treat you differently for asking. We may need to verify your identity first.

If you are in the UK or the EEA

You may lodge a complaint with your local supervisory authority — in the UK, the Information Commissioner’s Office at ico.org.uk. We would appreciate the chance to resolve it first.

If you are in California

Under the CCPA as amended by the CPRA, you have the rights above plus the right to know the categories of personal information we collect, the sources, the purpose, and the categories of third parties we disclose to — all set out in sections 2, 3 and 5.

In the last 12 months we have collected these categories: identifiers (name, email, IP address, browser identifier, device fingerprint, account and social identifiers, ad click identifiers, and the support chat’s browser and device identifiers); commercial information (subscription, plan and transaction records); internet and network activity (session and error logs, the ad and campaign you arrived from, and the pages the support chat records you visiting); audio and visual information (the videos, audio and images you bring, and the clips made from them); and inferences drawn from your transcripts for selecting clips and for improving our models. We collect them from you, from a social login provider if you use one, from the platform a video is imported from, from the link of the ad that brought you, and from Stripe.

You may use an authorised agent to make a request, with written proof of authorisation. You may also appeal a decision by replying to it.

If you are in Canada

PIPEDA gives you the rights above. If you are not satisfied with how we have handled a concern, you may complain to the Office of the Privacy Commissioner of Canada at priv.gc.ca.

If you are in Australia

The Australian Privacy Principles give you the rights above. If you are not satisfied with our response, you may complain to the Office of the Australian Information Commissioner at oaic.gov.au.

12. Children

The Service is not for children. You must be at least 16 to hold an account, and we do not knowingly collect personal information from anyone younger.

If you believe a child has given us personal information, tell us at support@kondense.app and we will delete it.

13. Changes to this policy

We will update this policy as the Service changes. The date at the top always shows the current version. If a change materially affects how we handle your information, we will tell you by email or in the app before it takes effect.

14. Contact us

For any privacy question, request or complaint: support@kondense.app.

Simon Chawla, carrying on business as Littlefoot Digital

242 Bryanston, Lucasville NS B4B 0N2, Canada

Nova Scotia, Canada