Jobtiv Sub-processor and Recipient List
Version 1.0
Last updated: [INSERT PUBLICATION DATE]
How to read this list
This page names the companies that handle personal information for Jobtiv, what each one does, what it can see, and where it is. Where a company we work with uses artificial intelligence providers of its own, we note that on its row. We review this page whenever we change a vendor, and at least once a year. If you think a company is missing, email support@jobtiv.ai.
It is divided into two parts.
Part A lists the companies that handle personal information on our instructions, under a contract that requires them to keep your information secure, to use it only to provide their service to us, and to delete or return it when we ask. The transfer route column records the safeguard we rely on when information goes outside the United Kingdom. Where a row says a position is under review, the note explains what that means for you and what to do if you would rather not use the feature. If you want the detail of the arrangements with any company on this page, ask us and we will tell you.
Part B lists other recipients. These are companies that decide for themselves how they use at least some of the information involved, so they are not simply acting on our instructions. We name them because you are entitled to know who is involved, even though their own privacy notices govern what they do.
"UK Extension" means the UK Extension to the EU-US Data Privacy Framework. "UK Addendum" means the International Data Transfer Addendum to the EU Standard Contractual Clauses issued by the Information Commissioner. "IDTA" means the International Data Transfer Agreement issued by the Information Commissioner. "Adequacy" means the country is covered by UK adequacy regulations, so no further safeguard is needed. Section 10 of our Privacy Policy explains what these are.
Part A
Core infrastructure
| Company and entity | What it does for us | What it can see | Where | Transfer route |
|---|---|---|---|---|
| Supabase (Supabase Pte. Ltd., Singapore) | Our database, sign-in system and file storage | Everything you store on Jobtiv: your account, CVs, documents, tracker, offers, uploaded images and files | London, United Kingdom. Your data is stored in the United Kingdom. Supabase itself is a Singapore company, so its own staff may access the system from outside the United Kingdom for support | UK Addendum |
| Vercel (Vercel Inc., United States) | Hosts and serves the website and our server code | Our server code: whatever a request carries while it is being handled, which for our tools includes your documents in transit. Vercel does not hold a copy once the request finishes. The routing layer: web addresses, request headers, your sign-in cookie, the guest identifier and a portfolio access key where one is used. It does not read request bodies, so documents do not pass through it. Two portfolio preview image features also run there and do handle portfolio content. Vercel keeps a record of requests for 30 days, which we have chosen deliberately rather than accepting a shorter default, so that we can investigate a problem properly | Our server code runs in London. The routing layer runs on Vercel's worldwide network and cannot be confined to one country. Vercel Inc. is a United States company, so its staff may access the platform from outside the United Kingdom | UK IDTA, deemed entered into under Schedule 3 of Vercel's data processing addendum. Vercel also states that it is certified under the EU-US Data Privacy Framework. We do not rely on that certification for United Kingdom transfers, because the Framework's UK Extension requires separate registration and we have not confirmed one |
| Upstash (Upstash, Inc., Delaware, United States) | Short lived caching and usage limits, using Redis; a small number of vector stores used to match roles and skills | Yes, CV content. A parsed CV is cached for up to 7 days against a fingerprint of the file, not against your account. The result of a CV scoring review is cached for 7 days in a separate database, also against a fingerprint of the content. A CV being turned into a PDF is held for 30 minutes against a one-time token | European Union and United Kingdom. Our Redis databases are in London and our vector indexes are in Ireland | UK Addendum for the company itself, which is in the United States. Upstash's own sub-processor list names Anthropic and OpenAI for generative artificial intelligence services. We review this page whenever we change a vendor, and at least annually |
| QStash (Upstash, Inc.) | Queues PDF generation jobs | Your account identifier, a job identifier and a filename. Not the CV itself, which sits in the cache above | As above | UK Addendum |
| Browserless (Browserless, Inc., United States) | Renders your CV to PDF in a headless browser | Yes, your fully rendered CV, for the duration of the render | London. We use their London endpoint (production-lon.browserless.io), so the rendering happens in the United Kingdom. Browserless, Inc. is a United States company | UK Addendum to the EU Standard Contractual Clauses, incorporated in their data processing addendum. See note 1 |
Artificial intelligence providers
| Company and entity | What it does for us | What it can see | Where | Transfer route |
|---|---|---|---|---|
| Anthropic (Anthropic Ireland, Limited, for UK customers) | Negotiation scripts, job description translation, cover letters, STAR answers, recruiter outreach drafts, portfolio copy, case practice, Application Tracker smart paste | Yes, CV content, and salary and offer figures | Anthropic Ireland, Limited contracts with us. Anthropic has not published the region in which its first party model interface processes requests, and we are asking. Assume the United States | UK Addendum. Anthropic commits contractually not to train its models on customer content. Retention on our account is set to 30 days |
| OpenAI | Fallback for several tools, Course Finder, voice transcription | Yes, CV content, and voice recordings | United States and other regions OpenAI operates | UK Addendum. OpenAI states that content sent through its developer interface is not used to train its models unless the customer opts in, which we have not. Request logging on our account is switched off. See note 2 on which OpenAI entity contracts with us |
| Google (Google LLC and its affiliates) | CV parsing, setup assistance, Studio rewrites, networking messages, portfolio fallback, course finding, voice transcription | Yes, CV content, and voice recordings | United States and other Google regions | We use Google's paid developer interface, so Google's data processing addendum applies and Google does not use what we send to improve its own products. We are confirming which Google entity contracts with us and which United Kingdom transfer safeguard applies, and we will publish the answer here |
| OpenRouter (OpenRouter, Inc., United States) | Routes requests for CV scoring review, Career Path Map, the cover letter editor and case and aptitude practice to a model operated by a further company | Yes, CV content, and salary figures on one fallback path | United States, with in-region routing available | Routing to model providers that train on submitted content is switched off on our account, and prompt logging is off. See note 3 |
| LlamaParse (LlamaIndex, Inc., United States) | Third route for parsing a difficult PDF | Yes, the entire PDF file you uploaded | United States. We use LlamaIndex's North America endpoint | See note 4. LlamaIndex states that uploaded documents are used only to return the result and never for model training |
Everything else
| Company and entity | What it does for us | What it can see | Where | Transfer route |
|---|---|---|---|---|
| Resend (Plus Five Five, Inc., United States) | Sends our email. At the moment that is sign-in messages, purchase and cancellation confirmations, renewal notices, security alerts and the feedback you send us. We do not send marketing email | Your email address, the content of the message, and for feedback the text you wrote and the page you were on | United States | UK Addendum. Resend's own sub-processor list names Anthropic and RunPod. We review this page whenever we change a vendor, and at least annually |
| Twilio (Twilio Inc., United States) | Sends the SMS code when phone verification is switched on. It is switched off at the moment | Your phone number | United States | UK Addendum, and Twilio's Binding Corporate Rules. Twilio is certified under the UK Extension. Twilio's own sub-processor list names Anthropic and Meta |
| PostHog (PostHog, Inc., United States) | Product analytics, where you have consented | Which features you use, your account identifier, your browser and approximate location. Our code is written so that the text of your CV, your messages and your documents does not go into an analytics event. We audit that rather than guarantee it, and section 4.2 of our Privacy Policy explains why we put it that way | Our account is on PostHog's European Union cloud | UK Addendum |
| Sentry (Functional Software, Inc., United States) | Error reports | The type of error, the code path, an anonymous account identifier. Request bodies, headers, cookies, query strings and breadcrumb contents are removed before the report leaves our servers | Germany. Our Sentry organisation is on their European region | UK Addendum. Sentry states it complies with the UK Extension |
| LangSmith (LangChain, Inc., United States) | Observability for our artificial intelligence calls. It receives nothing from the live Service. Tracing is switched off in our production code and the production project has never received any data. We use it while building, on a developer's own machine | Nothing in production. In development, whatever the developer was testing with | European Union endpoint | See note 9 |
| Slack (Slack Technologies Limited, Ireland) | Delivers operational alerts into our own workspace, such as a warning that our artificial intelligence spending has jumped | An internal reference and a number. It does not receive your documents | Ireland, with onward transfer to the United States | Adequacy for the transfer we make, because Slack Technologies Limited is in Ireland. Salesforce, which owns Slack, applies its own standard contractual clauses to onward transfers out of the European Union |
Part B: other recipients
These companies are not simply following our instructions. Their own privacy notices apply to what they do.
| Company | Why it is involved | What it receives | Its role | Transfer route |
|---|---|---|---|---|
| Stripe (Stripe Payments Europe, Limited, Ireland, and Stripe, LLC, United States) | Takes payment and manages subscriptions | Your email address and an account reference. Your card details, which go to Stripe and never to us. Not your CV or documents | Our processor for parts of the payment, and a controller in its own right for fraud prevention and its own regulatory duties | Adequacy for the Irish entity. UK Addendum, and the UK Extension, for onward transfer to Stripe, LLC. Stripe is certified under the UK Extension |
| Apollo.io (ZenLeads, Inc., United States) | Company and organisation lookup for Find Recruiters | A company name or domain. We search for organisations, not for named individuals. No CV content is sent | Controller of its own business contact database | We could not find Apollo on the Data Privacy Framework list. We are putting standard contractual clauses with the United Kingdom addendum in place. See note 8 |
| Hunter (Hunter Web Services, Inc., United States, hosted in Belgium) | Checks whether a company has a recruiting presence on a domain | A company domain | Controller of its own contact database | UK Addendum. Hunter's own sub-processor list names OpenAI and the Google Gemini API |
| Adzuna (Adzuna Ltd, United Kingdom) | Job market and salary data | A role and a location | Controller of its own job listings database | United Kingdom. No transfer. No signed data processing contract. See note 5 |
| Jooble (Ladoburn Europe Ltd, Cyprus) | Job market data | A role and a location | Controller of its own job listings database | Adequacy for the Cypriot entity. See note 5 on the Ukrainian operating entity named in Jooble's terms |
| Serper (operator not disclosed) | Web search for company research and Course Finder | Nothing about you. A generic search phrase such as a skill and a job title, or a company name. Never your CV, your name, your email address or an employer taken from your document | Not disclosed | None needed. We send no personal data to Serper, so there is nothing to transfer. See note 6 |
| Brandfetch (Brandfetch Sàrl, Switzerland) | Company logos | A company name or domain | Controller of its own request logs | Adequacy. Switzerland is covered by United Kingdom adequacy regulations |
| Logo.dev (Logodev, Inc., Delaware, United States) | Company logos | A company domain, and it logs the requesting IP address and timestamp | Controller of its own request logs | No executed data processing contract. See note 4 |
| Google (YouTube) and Udemy | Course search results | A search term | Controllers of their own platforms | Google's and Udemy's own published terms |
Notes
1. Browserless. Your CV passes through Browserless's systems while a PDF is being produced, and it is the only company on this page that sees your CV fully rendered. Two things have improved this position. We now use their London endpoint (production-lon.browserless.io), so the rendering happens in the United Kingdom rather than the United States. The name london.browserless.io was a vendor first answer and is not a production endpoint. And their data processing addendum carries the standard clauses the Information Commissioner has approved for transfers.
One thing we cannot yet tell you, and would rather say so than imply otherwise. Their addendum sets no limit on how long page content is held. We have asked, and we are waiting for an answer we can rely on. Until we have it, we do not make any claim about how long your rendered CV exists on their systems. We are separately building a route that produces the PDF in your own browser, which would remove the question. We are separately building a route that produces the PDF in your own browser, so that in time most downloads will not leave our systems at all.
2. OpenAI entity. OpenAI's published documents are not consistent about which OpenAI entity contracts with a United Kingdom customer. We have asked for written confirmation and will update this page when we have it.
3. OpenRouter. OpenRouter routes a request to a model operated by a further company. We have configured our account so that prompts and responses are not logged and so that requests are not routed to providers that train on submitted content. OpenRouter's own documentation makes clear that this setting governs routing rather than every upstream provider's own policy, so we cannot give an absolute assurance about every model that may serve a request. Section 7.1 of our Privacy Policy lists the four features that use it, so you can avoid them if you would rather.
4. LlamaParse. LlamaIndex makes its data processing agreement available on execution rather than incorporating it by default, and we are executing it. LlamaParse is our third parsing route and is only reached for a PDF the first two routes cannot read.
5. Adzuna and Jooble. Both return job listings. What we send is a search term, a role or a location, not your CV and not your name. Jooble's terms name an operating entity in Ukraine, which is not covered by United Kingdom adequacy regulations, and we are confirming in writing where the processing actually takes place. We will remove the integration if we cannot get a satisfactory answer.
6. Serper. Serper runs a web search for us. We have traced every place in our code that calls it and nothing we send identifies you: a generic phrase such as a skill and a job title capped at 80 characters, or the name of a company you picked. Your CV, your name, your email address and any employer named in your document stay with us. Because we send no personal data, Serper is not handling anything on your behalf and no information about you leaves the United Kingdom through it. Results come back from Serper's own index, and those results sometimes name a person, which section 4.5 of our Privacy Policy explains.
Serper does not publish its legal entity, its data location or its retention period, and has told us it will not enter into a data processing agreement below a spend level far above ours. We have kept the integration on the basis of the analysis above and we can switch it off with a single setting if that analysis ever stops holding.
7. Apollo. Apollo is not on the Data Privacy Framework list, so we do not rely on that route for it. What we send Apollo is a company name or a domain, not your CV and not your name. One honest qualification: the company box is a free text field, so whatever you type in it is what we search for. If you type a person's name there, that is what reaches Apollo, as an organisation search. We do not read it, classify it or attach it to you, and no identifier of yours travels with it.
9. LangSmith. We use LangSmith to see how our artificial intelligence calls behave while we are building. It is switched off in the live Service, in code rather than by a setting, and the production project has never received anything. Traces from a developer's own machine may include whatever that developer was testing with. If we ever switch it on in the live Service it will receive the text of prompts, which means your documents, and we will change this page and section 7 of our Privacy Policy before we do.
Changes to this list
We will update this page when a company on it changes, when we add one, or when we remove one. The date at the top shows the current version.
This page carries a dated change log at the bottom, updated whenever a company on it changes. We do not run a mailing list, because we would rather keep one page reliably current than maintain a list badly.
Change log
| Date | Change |
|---|---|
| [PUBLICATION DATE] | First published |
| 29 September 2026 | Browserless London host named: production-lon.browserless.io. The name london.browserless.io was a vendor first answer and is not production. |
Questions
Email support@jobtiv.ai. If you would like to see the transfer safeguards we rely on for a particular company, ask and we will send you a copy.
JOBTIV LTD, registered in England and Wales, company number 17092844. Registered office: 124-128 City Road, London, EC1V 2NX.
The other documents
These five documents work together, and each is a page on this site.
If any of these links does not work, tell us at support@jobtiv.ai and we will fix it. A legal page that does not load is a problem whatever it says.