Last updated: 18 November 2026
This privacy policy applies to Ergora Ltd, a company registered in England & Wales, trading as Ergora. We operate the marketing site at ergora.app, the product portal at ergora.cloud, and the cold-outreach domain try-ergora.com. For the purposes of UK GDPR and the Data Protection Act 2018 we are the data controller for personal data we process about our customers, prospects and website visitors.
You can contact us about anything in this policy, including to exercise your rights, at ops@ergora.cloud. Our data protection contact is reachable at the same address.
When you create an account we collect your name, email address, and (where set) a password hash. If you sign in via a social provider we receive your name, email and profile photo from that provider. We use this to authenticate you, contact you about the account, and personalise the experience.
You may add your company name, website, industry, brand assets, and connect third-party integrations (Shopify, Google, Meta, HubSpot, Xero and others). When you do, we store the resulting access tokens encrypted and use them only to retrieve data on your behalf within the scope you approved.
Google is connected one feature at a time, and Google asks you to approve each one separately. We only request the access that feature needs:
Most Google data is fetched live when you ask for it and is not stored. We keep only what a feature needs to work: monthly traffic and advertising totals for forecasts, keyword positions in the Rank Tracker and, if you use the Smart Inbox, the emails it has synced. Access tokens are encrypted at rest. When you ask a question that needs this data, the relevant part is sent to the AI model that writes the answer, only to produce that answer. That model is always Google’s Gemini on Google Cloud or Anthropic’s Claude, and neither provider trains its models on the data we send.
We do not sell Google user data, use it for advertising, or use it to train or improve AI or machine learning models. Nobody at Ergora reads it unless you ask us to for support, or it is needed for security or to comply with the law. You can disconnect Google at any time in Settings, Integrations, which deletes the stored tokens.
Ergora’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.
We process the prompts, files, voice notes, meeting recordings and other content you submit so the AI can respond. Voice and meeting audio is transcribed by Google’s Gemini model on Google Cloud. If that fails, the audio may be sent to OpenAI (Whisper) or AssemblyAI instead, as a backup. Audio you send only to be transcribed, such as a voice note, is not stored: the transcript is what we keep. Recordings of calls you make or upload in Ergora are different: the recording is kept in your account until you delete the call, which deletes the audio, the transcript, the recap and the notes it added to your CRM.
The product builds two kinds of memory from your interactions: your personal memory, which only you can see, and your business’s shared company memory. These memories belong to your workspace and are never shown to another customer. You can wipe your personal memory at any time, and correct or remove anything in your company memory. The one use of them beyond your own work is the Network Brain, below.
What: businesses on Ergora can be grouped into segments of similar businesses, such as the same industry. For each segment with at least five businesses, we send Google’s Gemini model on Google Cloud (Vertex AI) short extracts of each business’s company memory: the first 800 characters of its voice page, the first 600 characters of its offering page, and the number of studio outputs it shipped in the last 30 days. Each business is labelled with a number (B1, B2 and so on) instead of its account or project ID, although the extracts themselves may mention the business. Only those two extracts and that count are used: the Network Brain does not read your connected accounts itself, although your company memory can include facts Ergora has learned from them, such as your best-selling products.
Why: to find patterns that hold across the segment, such as how similar businesses describe their offer, and show them to the businesses in that segment in their Daily Briefing, so each can see what tends to work for businesses like it.
Safeguards: nothing is pooled in a segment of fewer than five businesses. The model is instructed to return only statistical or structural patterns, never to name a business, and never to include customer data, financial figures, contact details or copied text; this is an instruction to the model, and its output is not checked separately. Other customers see only the resulting patterns, never your company memory. Each pattern is set to expire 60 days after it is found, and is withdrawn at the first recompute after that. There is not yet a setting to switch this off. If you would like your business left out, ask us at ops@ergora.cloud and we will exclude it from the next recompute on; it then no longer sees the patterns either.
We collect credit use, images generated, feature usage, IP address, browser, device, session timestamps and similar technical data to operate the platform, enforce plan limits, prevent abuse and improve the Service. Stripe holds your payment-method details. We never see full card numbers.
If you ask for a free website read on ergora.app, we fetch the public homepage of the website you give (and its robots.txt, which we follow) and email the read to the address you gave, once. The read is put together from a fixed set of observations about the page, written by us, with figures our code works out from the page itself. An AI model from the providers listed below helps choose which of those observations apply, and picks from fixed lists what kind of business the page shows and who it seems to speak to; it writes no sentence of the email. Only when your email address is at the website’s own domain may the read also use two short phrases from the model (what the business offers and who it speaks to), each checked before it is used. The email has a link in it to stop any more reads being sent to your address. We send no other emails about it.
We store your request (the email address, the website address, the page you asked from, a keyed hash of your IP address, which helps prevent abuse of the form and cannot be turned back into the address without our key) and the read. We keep your request and the read for no more than 90 days, then delete them. Two things are kept longer, and neither holds your email address or your IP address in a form anyone can read: if you use the stop link, a scrambled record of your address, kept so that it is never sent a read again; and, for 8 days, scrambled counters that hold the form’s limits.
The free tools on ergora.app ask for an email address in one place: the full-size downloads of the Product Photo Background Remover. If you leave your address there, we store it with the name of the tool and a keyed hash of your IP address, which helps prevent abuse of the form. We send you no emails. We keep your address for no more than 90 days, then delete it.
The website read form and the free tools’ email gate check for bots on our own server, with a hidden field and a check of how long the form took to fill in. If we switch on a further check, Cloudflare Turnstile, it loads from Cloudflare only once you start using one of those forms, never before, and Cloudflare receives your IP address and information about your browser and device to tell people from bots. It is a security measure, not analytics or advertising. When it is switched off, nothing from Cloudflare loads.
If you choose “Accept all” in our cookie banner, on ergora.app or in the portal, Google Analytics 4 sets cookies that record usage patterns, and we enable Google Signals and Google’s advertising cookies so that Google can show you Ergora ads on other websites based on your visit. On ergora.app, and in the portal at ergora.cloud whether you are signed in or not, the Google tag is not loaded at all, and no analytics or advertising data is sent to Google, until you choose “Accept all”. If you later withdraw it, no more analytics or advertising data is sent to Google from that site, and the Google Analytics cookies on it are removed. The full list, and how to change your mind, is in our Cookie Policy.
If you send us a message through the contact form or register interest in a product, we store what you enter (your name, email, company, team size, the specialists you ticked and your message) and add you as a contact in Loops, our email provider, so we can reply and follow up on that enquiry. Your cookie choice is stored in your browser together with a random identifier, which is sent to ergora.cloud with your choice so the portal can respect it too.
We also contact businesses we think Ergora could help, from try-ergora.com, on the basis of legitimate interests. We find them, and the email address they publish, from public sources: directories of online stores and of professional firms, lists of popular websites, and the businesses’ own websites. If you arrive on ergora.app from one of those emails, the link carries a tracking code, and we record that you arrived (the code, the page, any campaign tags, the referring site, your browser type and a hashed form of your IP address) so we know which emails are useful. You can ask us to stop at any time, by replying or by writing to ops@ergora.cloud.
We process personal data on the following lawful bases under UK GDPR: performance of a contract (operating your account, delivering the Service, processing payments), legitimate interests (securing the Service, preventing fraud, improving the product, and reaching out to business prospects in line with PECR), consent (where required, for example for non-essential cookies and marketing email), and legal obligation (for example, retaining tax records).
We share data only with the third-party processors that help us run the Service, with integration providers you choose to connect, and where we are required to by law. We do not sell your data. The only advertising-related sharing is with Google, and only if you choose “Accept all” in our cookie banner (see Analytics and advertising cookies above).
| Processor | Purpose | What it receives | Region |
|---|---|---|---|
| Supabase | Database, authentication and file storage | Account info, project data, files, embeddings, OAuth tokens | EU (Frankfurt), primary |
| Stripe | Payment processing and subscription management | Name, email, billing address, card data (held by Stripe, not us) | US / EU |
| Loops | Transactional and lifecycle email | Email address, name, account events | US |
| Google Cloud (Vertex AI) | Our main AI models: Gemini for chat, answers and reports, Gemini and Veo for images and video, plus speech-to-text, text-to-speech and knowledge search. Google does not use this data to train its models | Prompts and the context sent with them, including data from your Google account when you ask about it, attached files, voice notes and meeting recordings, and for the Network Brain, short extracts of company memory | US (us-central1). Gemini 3 models and speech-to-text run on Google’s global service, which may process a request in any Google Cloud region |
| Anthropic | Claude, for the most demanding tasks such as Studio's Design layouts, meeting summaries and the AI Visibility audit. Anthropic does not use this data to train its models | Prompts and the context sent with them | Stored in the US. A request may be processed in any region where Anthropic runs its models |
| DeepSeek | A fast, low-cost model for bulk work such as sorting, scoring and short drafts, and a backup for chat. Its terms let it use de-identified inputs to improve its services, so we never send it data from your Google account | Prompts and the context sent with them | China |
| Kling AI | Video generation in Studio's AI video tab. Kling does not use this data to train its models | Video prompts and the images or clips you add | Singapore |
| xAI | Searches X for posts that mention your brand or the topics you track, using Grok. xAI does not use this data to train its models | The brand names, handles and keywords you ask us to monitor | US |
| Browser Use | Only when you switch on Stealth mode in the built-in browser: runs that session on Browser Use’s cloud browsers | The pages you open and anything you type in that session | Mainly US (Amazon Web Services) |
| OpenRouter | A reserve route to other AI models if our direct providers are unavailable, and the Perplexity check in the AI Visibility audit. OpenRouter itself does not train on prompts | Prompts and the context sent with them, never data from your Google account | US |
| OpenAI | A backup for transcribing voice notes and meeting recordings (Whisper) if Google’s transcription fails. Also, only if you add your own OpenAI key to the website chat widget, it answers your visitors’ questions under your own OpenAI account | Audio to transcribe; for the widget with your own key, the visitor’s question and your widget’s FAQ and business context | US |
| AssemblyAI | A further backup for transcribing voice notes and meeting recordings if the services above fail | Audio to transcribe | US |
| Jina AI | A last-resort backup for knowledge search: turns a search query into an embedding if our main embedding services are unavailable | The text of the search query | Jina AI GmbH, Germany |
| Hostinger | Server hosting for the portal, the marketing site and supporting services, including the video-call server for Ergora calls | Encrypted application data and logs; the audio and video of Ergora calls pass through it | UK / EU |
| Sentry | Error monitoring for the portal: when something breaks in your browser or on our servers, a report of the error is sent so we can fix it. Session replay is switched off | The error, the page it happened on, your browser and device type, and your IP address | US or EU (Sentry keeps data in one of the two) |
| ip-api.com | Only for Ergora Remote, which is not on sale: finds the city and country of the IP address a device registers from, so an unfamiliar location can be checked | The device’s IP address | Not published by the provider |
| Public data lookups | When you add a contact or keep mentioning a company, we look up public facts about it: company filings (Companies House), area figures for a US ZIP code (US Census), news (GNews, Currents), the location of an IP address the contact came with (ipapi.com) and public holidays (Nager.Date) | A company name or email domain, a ZIP code, a country, and a contact’s IP address if one was recorded | UK (Companies House), US (US Census); the others do not publish a single region |
| DataForSEO | Search volumes for the keywords you research in the SEO tools | The keyword and the country you research it for | Not published by the provider |
| Cloudflare (Turnstile) | Only if we switch on the security check on the website read form and the free tools’ email gate: tells people from bots. It loads only once you start using one of those forms | Your IP address, and information about your browser and device and how the check was completed | US (Cloudflare runs the check on its global network) |
| Oxylabs | A fallback for fetching public web pages you ask Ergora to read (a site to audit, a search result to check) when our own fetcher cannot reach them | The address of the public page and what it returns | EU (Lithuania) |
Where a processor offers data processing terms, they require it to keep your data confidential and secure and to use it only to provide its service to us. The table notes where a provider’s own terms allow more than that.
Our database is in the EU (Frankfurt) and our servers are in the UK and EU, but most AI processing happens in the United States, as the table above shows. The processors that handle data in other countries are in the United States (Stripe, Loops, Google Cloud, Anthropic, OpenAI, AssemblyAI, xAI, Browser Use, OpenRouter, Sentry, US Census and, if the forms’ security check is switched on, Cloudflare), Singapore (Kling) and China (DeepSeek). Google’s global AI service and Anthropic may also process a request in other regions where they run their models. Singapore and China do not have a UK adequacy decision. Jina AI, our last-resort backup for knowledge search, is based in Germany; where it processes a query is available on request. Where a processor offers them, we rely on the UK-US Data Bridge, the EU-US Data Privacy Framework, or Standard Contractual Clauses with the UK Addendum to keep your data protected to UK GDPR standards. You can ask us for details of the safeguards for any processor at ops@ergora.cloud.
We keep your account data for as long as your account is active. If you delete your account in Settings, any subscription is cancelled and the account is deleted at once, with the data linked to it in our database. If you ask us to delete it instead, we do so within 30 days. Either way, we keep what we are required to keep for longer (for example, billing records that we must retain for tax purposes for up to 7 years).
Our web servers keep a standard log of requests (the page requested, the time, the browser type and the IP address) for 30 days, to keep the Service secure and to count visits. It is then deleted.
While your account is active, the portal also clears some things on a fixed schedule, so it keeps only what it needs:
Backups are rotated on a 30-day cycle, so deleted data may persist in encrypted backups for up to 30 days after deletion before being permanently overwritten.
You have the right to access the personal data we hold about you, ask us to correct inaccurate data, request erasure, object to or restrict certain processing, withdraw consent where processing is based on it, and receive a copy of your data in a portable format.
You can exercise the most common rights yourself, instantly, from within the product:
GET https://ergora.cloud/api/account/export while signed in.POST https://ergora.cloud/api/account/delete while signed in.For anything else, email ops@ergora.cloud. We will respond within one month. If you are not satisfied with our response you have the right to lodge a complaint with the UK Information Commissioner’s Office (ico.org.uk).
Ergora Remote, a desktop agent that lets Ergora read files on your own computer, is not on sale. Where a device was registered to it during testing, we record the device and the city and country it registered from (never precise GPS coordinates), to spot an unfamiliar location and ask an approved device to confirm it. To find the city and country, we send the IP address to ip-api.com, which returns only the location. We keep it with the device’s registration and never sell it. Registered devices are included when you download your data.
We use a small number of essential cookies to keep you signed in and to remember your preferences, plus optional analytics and advertising cookies, which are set only if you choose “Accept all”. See our Cookie Policy for the full list, and to change your choice at any time.
The Service is not directed at children, and we do not knowingly collect personal data from anyone under the age of 16. If you believe a child has provided us with personal data, please contact us and we will delete it.
We use encryption in transit (TLS) and at rest, row-level security in our database, encrypted OAuth tokens, access-controlled production secrets, and least-privilege access for our team. No system is perfectly secure; if we ever suffer a breach affecting your personal data we will notify you and the ICO in line with our legal obligations.
We may update this policy from time to time. The “Last updated” date at the top of this page reflects the latest revision. Material changes will be communicated by email or an in-product notice in advance.
For privacy questions, data subject requests, or to contact our data protection lead, email ops@ergora.cloud.