Privacy Policy

NenoSource is run by Piyush Nakrani, an individual trading as NenoSource, based in Surat, India (“we”, “us”). There is no incorporated company behind the service — it is a sole proprietorship, and we would rather say so than imply otherwise. We decide what personal information is collected and why, which makes us the data controller for everything described here.

For anything in this policy — including a request to see, correct or delete your data, or to report a security problem — write to nenosource01@gmail.com. A person reads it.

When you create an account we store:

• Your email address, which identifies the account and is where we send verification, password-reset and workspace-invite messages.
• A hash of your password, produced with bcrypt. We do not store your password itself and cannot recover it — a reset replaces the hash.
• Profile details you choose to fill in: display name, username, short bio and an avatar image.
• Preferences: your light/dark theme choice, used only to render the interface.

When you sign up we also check the email address before accepting it: the domain is compared against a list of known disposable-mail providers, and we make a DNS lookup to confirm the domain can receive mail. The lookup asks a DNS resolver about the domain, not about you, and no part of your address is sent to a third party for this.

If you use “Continue with Google”, Google returns a signed identity token that we verify. From it we read your Google account identifier, email address, name and whether Google has verified that email. We store the identifier so the same Google account signs you back into the same workspace next time.

We do not receive your Google password, and we ask for no access to any other Google service — no Drive, no Gmail, no contacts.

The service answers questions from documents you provide. You can upload PDF, Word and plain text files, up to 25 MB each. We store the file itself, its file name, size, type and a checksum, and then process it: the text is extracted, split into passages, and each passage is converted into a numeric vector (an “embedding”) so that it can be searched by meaning rather than by keyword. The extracted passages and their vectors are stored alongside the file.

Documents belong to the workspace you uploaded them to, and are attached to a particular agentwithin it. A search only ever looks at the documents attached to the agent you asked, which means one workspace’s documents are never used to answer another’s questions — including across different customers — and one agent’s documents are not used to answer a different agent’s questions either.

Please do not upload documents containing information you are not permitted to share with a third-party AI provider. AI providers and what is sent to them explains exactly what leaves our systems.

We store your conversations: the questions you ask, the answers produced, and metadata about each message such as which model answered and how many tokens it used. Conversations belong to the account that created them and are not visible to other members of your workspace.

Citation snippets outlive the document they were taken from. Each answer cites its sources by storing a copy of the quoted passage, the source name and the page number, so that a past answer stays checkable. This copy is deliberately independent of the document: if you later delete a knowledge source, its passages and vectors are removed and it can no longer be used to answer anything, but the snippets already quoted in your message history remain there. To remove those as well, delete the conversations that contain them, or delete your account.

Paid plans can call an agent from your own software using an API key. We store nothing from those conversations — not the questions, not the answers, and nothing about the person who asked. They are answered and then gone.

What we do keep is one record per call, so that we can bill you and help you when something breaks: which key was used, which agent answered, how many tokens it took, whether it succeeded, and when. That record is about the request, not its contents.

If you build NenoSource into your own product, the people using that product are your users, not ours. We have no account for them, no relationship with them, and receive nothing about them beyond the text of the question you send us on their behalf. Explaining your own use of AI to them, and handling their data, is yours to do.

When you sign in we create a session and record, against it, the IP address and browser user-agent string the sign-in came from, together with the times the session was issued and last used. We use this to detect a stolen session token being replayed, and we keep it so that we can later show you a list of your own active sessions.

That list does not exist yet, so today your control over sessions is all-or-nothing: you can sign out of every device at once from the account menu, which you should do if you have signed in on a shared or lost device. There is currently no way to see your sessions individually or to end just one of them. Sessions also expire on their own.

We use no advertising cookies and nothing that follows you to other websites. There is no advertising pixel, no ad network, and no session recorder — we never record your screen. We do use one analytics tool, and only if you say yes: see product analytics below.

These three cookies are set for everyone, because the service cannot work without them:

In your browser’s storage we keep only the identifier of the workspace you last had open, so the app reopens where you left it. The short-lived token used to authorise API calls is held in memory and is gone when you close the tab.

Our servers also record that certain product actions happened, whether or not you accept analytics. An agent was created, a document finished indexing, a message was sent — so we can tell where the product is failing people. That record is made on our side while serving your request: it involves no cookie and stores nothing on your device, which is why it is not part of the choice below. It identifies you only by your account and workspace identifiers, and contains no document text, no messages and no email address. The provider that receives it is listed under service providers.

We would like to understand how the product is used — which pages people visit, where they get stuck, which features go untouched — so we can fix the parts that are failing. For that we use PostHog, and it does nothing at all unless you agree to it first.

The first time you visit we ask. Until you answer, and permanently if you decline, the analytics software is never even downloaded to your browser — no request is made to PostHog, no script runs and no cookie is set. This is not a notice about something already happening. Declining costs you nothing: every feature works identically either way.

If you agree, what is collected

• Pages you visit within this site, and the address of the page that linked you here.
• A small, fixed list of actions: starting a sign-up, viewing the pricing section, beginning a checkout.
• Ordinary technical details any web server sees — browser type, screen size, and your approximate location derived from your IP address. Your IP address is not stored alongside these events.
• Once you sign in, your account identifier, so we can tell that the person who arrived from a search is the same person who later signed up.

What is never collected, whatever you choose

We do not record your screen, and we do not capture your clicks automatically. Only the specific actions listed above are sent, and each one carries a fixed set of fields decided in advance. Your documents, the text of your conversations, your agents’ instructions, your name and your email address are never included.

Changing your mind

You can turn analytics off at any time under Settings → Appearance. Doing so stops collection immediately and deletes the cookies PostHog set. Because the choice is stored in your browser rather than on your account, it applies to the browser you made it in — if you use the product on a phone and a laptop, each will ask you once.

If you accept, PostHog sets its own cookies to recognise your browser between visits. They exist only while your answer is yes.

Where these events go

PostHog processes these events in the United States. If you are in the EEA or the UK, the basis for sending them there is set out under if you are in the EEA or the UK. Declining analytics means nothing is sent at all.

Running this service requires sending your content to an AI provider. We use OpenAI, for both answering questions and indexing documents. This happens at two separate moments, and they differ in how much is sent.

When you upload a document

The entire text of the document is sent to the embedding provider, not an extract. Indexing works by converting every passage of the document into a vector, so every passage has to be read. This happens once, when the document is first indexed, and again if it is re-indexed. If a document is too sensitive to share with a third-party AI provider in full, it is too sensitive to upload here.

When you ask a question

Three things are sent, and only these three:

1. The instructions you gave the agent.
2. The passages retrieved from your documents for that question — typically a handful, because only the closest matches are retrieved.
3. The earlier turns of that conversation, and your question.

What the provider does with it

We send this through OpenAI’s business API. We do not authorise them to train models on your content, and we do not sell your data or use it to train models of our own. Your account name, email address and billing details are never included in what we send.

The provider keeps a copy for a period of its own, and we do not control it. Their published API terms allow them to retain what we send for a limited window — commonly up to 30 days — for abuse monitoring, before deleting it. We have not negotiated zero-retention terms. That retention is governed by the provider’s policy rather than this one, and it applies to your document text and your questions alike.

We share personal information only with the providers that operate the service, and only what each one needs:

We may also disclose information where the law requires it, or to establish or defend a legal claim. If the business is ever sold or merged, your information may transfer with it; we will say so before that happens.

Paid plans are sold through Paddle, which acts as the merchant of record. You enter your card details on Paddle’s checkout, not on ours — we never see or store your card number.

What we keep is what we need to know what you are entitled to and to show you a receipt: your Paddle customer and subscription identifiers, your plan and its status, the date the current period ends, and the amount, currency, status and invoice link of each payment. We also retain the notifications Paddle sends us about your subscription, which form the record of what was charged and when.

Paddle handles your payment data under its own privacy policy, and is responsible for the sales tax and VAT/GST on your purchase.

We send only transactional email: confirm your address, reset your password, you have been invited to a workspace, and your password was changed. There is no marketing list, and we do not send promotional email.

Your account, documents, agents and conversations are kept for as long as your account exists. Beyond that:

• Unverified accounts are released automatically. If you never confirm your email address, we warn you and then delete the account, freeing the address for re-registration.
• Expired tokens — sign-in sessions, verification links, password-reset links and unaccepted workspace invitations — are purged automatically once they lapse.
• Payment and subscription records are kept after cancellation, because tax and accounting rules require us to retain them.
• Usage and cost records. We log each AI request with the model used, token counts, latency and cost, in order to meter plan limits and investigate faults. These records contain no document text or message content. We have not yet set a retention period for them, so at present they are kept indefinitely. We intend to define one; we would rather say this plainly than state a period we do not currently enforce.

You can delete your account from the settings page. Because it cannot be undone, we ask for your password and ask you to type the word DELETE to confirm.

Deletion is permanent, not a flag on a row we keep: your profile, conversations, messages and citation snippets are erased, and your avatar is removed from storage. We have no way to restore an account afterwards. There is no bulk export — documents can be downloaded one at a time from the agent that holds them, so retrieve anything you want to keep before you delete, and copy out any conversation you need. Records we are legally required to retain — chiefly payment records, as described above — survive deletion in the minimum form the law requires.

What happens to your workspaces depends on whether anyone else can run them. A workspace that has another owner, or that you are not an owner of, is not deleted — you are simply removed from it, and its agents and documents stay with the people who still use it. A workspace you are the only owner of is deleted along with your account, including everything inside it, because there would otherwise be nobody able to reach or remove it.

Before you confirm, we show you exactly which of your workspaces fall into each case and how many other people would lose access. If deleting your account would destroy a workspace that other people are members of, we ask you to confirm that specifically — and you can instead make someone else an owner first, which leaves the workspace running without you. If you want documents you uploaded removed from a workspace that will survive, delete them before you delete your account.

Any paid plan on a workspace that is deleted is cancelled immediately, and the remainder of the period already paid for is not refunded.

The usage and cost records described under how long we keep information also remain, because the workspace still has to account for what it spent. What is erased is the link between that spending and you: the records are detached from your conversations rather than kept alongside them.

Passwords are hashed with bcrypt. The cookie that keeps you signed in is HttpOnly, restricted to the authentication endpoints, and marked Secure in production; the token used for API calls is short-lived and never written to disk. Sessions can be revoked, and a reused session token invalidates the whole family it belongs to. Uploads are checked against their declared type by inspecting the file’s own bytes, not by trusting its extension. Every query for your documents carries your workspace as a condition, so isolation does not depend on a caller remembering to filter.

No service can promise perfect security, and we do not. If you find a vulnerability, please report it to nenosource01@gmail.com.

Under India’s Digital Personal Data Protection Act, 2023, you may ask us to confirm what personal data we hold about you and who it has been shared with, to correct or complete anything inaccurate, to erase data we no longer need, and to withdraw a consent you previously gave. You may also nominate someone to exercise these rights on your behalf if you die or become incapacitated.

Most of this is available to you directly in the app: your profile is editable, and documents, conversations and your account can all be deleted. Account deletion is immediate and irreversible, though it does not reach content owned by a shared workspace — deleting your account sets out exactly what it does and does not remove. For anything else, write to nenosource01@gmail.com and we will respond within the period the law allows. If you are not satisfied, you may complain to the Data Protection Board of India.

Where the EU or UK GDPR applies to you, we rely on these lawful bases: performance of a contract for operating your account and answering your questions; legitimate interests for keeping the service secure, preventing abuse and metering usage; legal obligation for tax and accounting records; and consent where we ask for it, which you may withdraw at any time.

You have the rights of access, rectification, erasure, restriction, portability and objection, and the right to lodge a complaint with your local supervisory authority.

On portability specifically, we would rather be accurate than reassuring: the app has no one-click export today. Documents can be downloaded individually, and for a machine-readable copy of the rest of your data you can ask us at nenosource01@gmail.com and we will produce one within the period the law allows. We are building the self-service version.

We are established outside the EEA and the UK, and the providers listed under service providersoperate internationally, so your data will be transferred across borders — including to India, where we are based, and to the United States, where our AI, storage and analytics providers process it. Those transfers are made under the European Commission’s Standard Contractual Clauses (and the UK Addendum where relevant), together with the safeguards in our agreements with each provider.

We have not appointed a representative in the EEA or the UK under Article 27. We are a one-person operation outside both, and we would rather tell you that than name one we do not have. Write to us directly at nenosource01@gmail.com and we will handle your request ourselves.

The service is not intended for children, and you must be at least 18 to open an account. We do not knowingly collect data from a child; if you believe we have, contact us and we will delete it.

If we change this policy we will update the date at the top of the page, and for any change that materially affects your rights we will tell you by email or in the app before it takes effect.