1. Who operates the platform
Flowcontent is a software-as-a-service platform operated by eozore, the company identified (tax ID and address) in the footer of every page on this site. It produces an article, a video, and social pieces from a single topic, and publishes the result to the customer’s own accounts, with authorization granted by them.
In LGPD terms: for the content produced and the connected accounts, the customer is the controller and Flowcontent is the processor — we handle that data on the customer’s instructions, to deliver the contracted service, and for nothing else. For the customer’s own account data (sign-up, subscription, usage), Flowcontent is the controller.
For anything related to privacy, including access or deletion requests, write to contato@eozore.com.
Data protection officer (LGPD art. 41): eozore itself, through contato@eozore.com. Data subjects and the Brazilian Data Protection Authority (ANPD) reach us there.
2. What data is processed
Four categories, with distinct purposes:
- Account. Email and name from the login provider (Google) or from email-and-password sign-up, plus the full name and date of birth you give at registration. Used to identify you, confirm you are an adult, apply your plan, and send operational notices.
- Subscription. Plan, payment status, billing identifiers, and the subscriber’s CPF or CNPJ (Brazilian tax ID), which the payment processor requires to issue charges and which we also use to stop one person from opening several free trials. The CPF is stored encrypted, compared only through an irreversible fingerprint, and never shown back in full. Card data never passes through here — payment is processed by Asaas, and the platform only receives confirmation that the subscription is active or not.
- Content and credentials. The topics, articles, scripts, and pieces you produce, and the credentials you connect: social network tokens and the keys to your own avatar and voice accounts.
- Image and voice. Photos of your face and your voice, when you turn on the generated avatar, and the character images you upload. They have rules of their own, in section 5.
No data is sold, there is no advertising profiling, and nothing you produce feeds model training.
3. How credentials are stored
The keys you bring (HeyGen, ElevenLabs) and the tokens of connected networks are encrypted before they reach the database, with AES-256-GCM and a key derived per workspace — so that one customer’s ciphertext is not decryptable in another’s context.
The platform never shows a credential back to you. Once saved, it can only be replaced or deleted. That is deliberate: a screen that displays the key turns any improper access to your session into a leak of the key.
Deleting a credential deletes the ciphertext. We keep no copy.
4. Data accessed through Google and YouTube APIs
When you connect a YouTube channel, Flowcontent acts on your behalf and with your authorization, to publish what you approved. The platform does not access the account of anyone who has not connected it.
Flowcontent uses YouTube API Services. By using the service you agree to the YouTube Terms of Service, and Google’s handling of data is governed by the Google Privacy Policy.
The permissions requested from Google, and what they are for:
youtube.upload— upload the videos you approved to your channel. Videos are uploaded as private; making them public is always your decision.youtube.readonly— confirm the authorization is still valid and identify the destination channel before uploading.yt-analytics.readonly— read the performance of your channel’s videos (views, watch time, retention, subscribers gained), to show you what worked and guide the next content. Read-only, your channel only. Requested only when metrics reading is available on your account; the numbers stay in your workspace and only enter statistics pooled across accounts if you chose to join the knowledge exchange.
The authorization token is stored encrypted and used only by this platform’s services, to carry out what you approved. It is not shared with third parties, not used for advertising, does not feed model training, and is not transferred to any other system.
Flowcontent’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.
Authorization can be revoked at any time, in the platform itself or at myaccount.google.com/permissions. Revocation takes effect immediately and prevents any further upload.
5. Image, face and voice
The generated avatar makes videos of you speaking from photos of your face and your voice. That data identifies a person, so it follows stricter rules than the rest:
- Only with your consent, given at upload. Photos and voice are used only to generate the videos you ask for, in your workspace. Nothing else: no profiling, no identification, and no model training.
- Only your face, or that of someone who consented. Uploading another person’s photo or voice requires their written consent. Character images you upload come with your statement that you hold the rights to them.
- Who processes it. Video generation runs on Google’s video models (Vertex AI) or on fal.ai, depending on the engine chosen; the switch to your voice runs on ElevenLabs, on your own account. Each receives only what it needs to generate each segment.
- Delete whenever you want. Each photo can be deleted under Account › Generated avatar, effective immediately for later videos. Without a request, photos are deleted along with the account (section 8).
The output — the videos with your avatar — is your content, and follows the content rules in section 2.
6. Other connected platforms
The same rules apply to LinkedIn and to the Meta family (Instagram, Threads, and Facebook): you make the connection, tokens are encrypted, nothing is shared, and authorization is revocable both here and in each platform’s settings.
On Meta, the instagram_manage_insights permission reads the performance of your posts (reach, impressions, interactions, saves), to show what worked and guide the next content. Read-only, only for the account you connected, and requested only when metrics reading is available on your account. When you remove the app on Facebook, Meta notifies us and we delete the connection (Data deletion page).
7. Subprocessors
To deliver the service, the platform relies on suppliers that also process data. They are, and what for:
- Google Cloud — hosting, database, and the artificial intelligence models that write and generate images and video, through Vertex AI (Google’s own models and partner models offered there). Data sent is not used to train those models.
- fal.ai — generation of avatar videos when the chosen engine runs there. It receives the photos and the text of each segment, only to generate it.
- Asaas — subscription processing. They handle payment data; we never receive it.
- Resend — transactional email delivery.
- HeyGen and ElevenLabs — avatar and voice, on your own account, with the keys you connected. Your contractual relationship with those two is direct.
- Pexels and Pixabay — free image and video libraries that supply the supporting shots. They receive only search terms, never your data.
8. Where data lives and for how long
Data is stored on Google Cloud Platform, on servers in the United States. The international transfer relies on the supplier’s standard clauses and on article 33 of the LGPD.
Your workspace’s content, credentials, and avatar photos are kept while the account exists. After cancelling, you may request deletion at any time; absent a request, data is deleted 90 days after termination. Records required by law (tax and access logs) follow their own statutory periods.
Record of model calls. We keep the content of each call to the artificial-intelligence models — what was sent and what came back — for up to 180 days, only to investigate problems and rerun a job; after that the content is deleted. Records that describe the call without the content (when, which model, how much text was processed, cost) are kept for 24 months. Photos of your face and your voice never enter this record. All of it is deleted along with the account.
10. Your rights
Under Brazil’s General Data Protection Law (LGPD, Law 13.709/2018) you may request confirmation of processing, access, correction, portability, anonymization, or deletion of your data, and withdraw consent.
Write to contato@eozore.com. Requests are answered within 15 days. The step-by-step for deletion — including for people who connected through Facebook or Instagram — is at Data deletion.
When a request comes from someone whose data we process as a processor — for example, someone appearing in content produced by a customer — we forward the request to the controlling customer and support them in answering it.
11. Incidents
In the event of a security incident that may pose material risk, we notify affected customers and the ANPD within the LGPD deadlines, with what is known about the event and the measures taken.
12. Knowledge exchange between accounts (optional)
It is not active yet. When it is, what is written here applies — and you turn it on in a screen that repeats this exchange in three lists, with no fine print.
- What you donate: the performance of your pieces relative to your own average (for example, "1.4× the median"), tied to the technique used, the format, the network and the niche. Never the absolute value.
- What you receive: the network’s learnings by niche, format and technique (for example, "in your niche, a number hook held 1.4× the median across 9 accounts").
- What never leaves your account: text, script, video, image, face, voice, brand, name, what the CMO knows about your business, absolute numbers (spend, revenue, followers) and your customers’ data.
Nobody is identifiable. A learning is only published when at least 5 accounts contribute to it, and only anonymized aggregates circulate — which, because they identify no one, are not personal data.
3-month minimum. Whoever turns the exchange on stays in for at least 3 months; otherwise someone would turn it on, use what the network knows and turn it off the following week. After that period you can turn it off whenever you like, and donating and receiving stop at the same moment.
The LGPD applies on top of the minimum. The right to withdraw consent has no deadline: if you ask to stop before 3 months, we stop — and access to the network’s knowledge goes with it. Turning it off does not undo the past: what already entered an anonymous aggregate does not leave it, because there is no way to know whose it is.
This does not change the rest of this policy: your account’s content and data still do not train any model. Data-subject requests (access, correction, deletion) are handled through the contact below, under Your rights.
13. Changes
Changes to this policy appear on this page, with the date at the top revised. Material changes to how personal data is handled are announced before taking effect.