Legal

Version 2026-11-04 · Effective 4 November 2026

Terms of Service

These terms govern your use of the service operated by Exon Biosciences (“we”, “us”). By creating an account you agree to them. If you are agreeing on behalf of an organisation, you confirm you are authorised to bind that organisation.

The service

We provide a semantic layer that connects to data sources you authorise, maps them to a shared schema, and makes the result queryable through dashboards, BI tools and automated agents. You keep ownership of your data and of the credentials you supply.

Your responsibilities

You are responsible for having the right to connect the data sources you connect, for the accuracy of that data, and for keeping your account credentials secure. You must not use the service to process data you are not permitted to process, or to attempt to access another customer’s data.

Billing

Seats are billed per user, monthly or yearly, and renew automatically until you cancel. Separately from seats, the service charges for: building a semantic layer over a connected source, charged once for each column mapped on a connection (a column whose values are split into several facts, such as a product name that includes its supplier, is still one column); mapping a column again because its source changed — when a column changes type, or is renamed in a way the service can recognise only by mapping it again — charged for each such column, once per change; on the Pro plan, an annual platform fee per mapped column for keeping that layer live, subject to a minimum per connection per year; AI assistant access, per person given it, per month; and data migration — for each migration run, a base fee that covers a first block of rows plus a rate for each row beyond it, and for each cycle of a recurring sync, a rate for each row the source has gained beyond the most it held at an earlier cycle.

A column is charged once per connection, not once per source: a source you delete and connect again, or a file read by the desktop agent that you move or rename, becomes a new connection, and its columns are charged again. When a source stops providing a mapped column, the column is archived, not deleted: it is still counted as a column you have paid for, so a column the source adds later is a new column and is charged as one, unless the service recognises it as the same column renamed or moved to another table, which it then follows instead. An archived column that comes back is restored at no charge. A migration run is not charged when the build fee for its source was paid for at least half of that source’s mapped columns (columns covered by the free allowance are not paid columns). On the Enterprise plan a migration run has no base fee: rows are included up to an allowance over each rolling twelve months, and each row beyond it is charged. Neither of these applies to the cycles of a recurring sync.

AI work on your semantic layer is included, not metered. Mapping your columns, recognising the same company, product or person across your sources, reconciling duplicates and checking data quality are paid for by the build fee and the platform fee, and are never charged for by how much AI they use. The one exception is mapping a column again because its source changed, described above, which is charged per column. Following a source change without a new mapping — archiving a column, restoring one, or following a rename or a moved table the service recognises by its name or its values — is not charged. Data connections are not charged for separately.

How the charges other than seats are raised. They are raised by the service at the rates published for your plan at the time, and not every one is quoted to you at the moment it is incurred:

• The build fee is invoiced at the next billing pass for the columns mapped since the previous one. When you set a source up from its connection page, its estimated price is shown before you start. Columns mapped without that step — by Quick Setup, when the desktop agent picks up a new file or a new column, or when automatic sync finds a new column in a source — are billed the same way, without a separate quote.
• Mapping a column again after a source change is invoiced at the next billing pass, with the build fee, without a separate quote: every mapping run handles the source changes it finds, whether automatic sync, Quick Setup, the desktop agent or you started it. A change the service cannot yet confirm — for an API, a field missing only from its newest records, or a type read from too few of them — is not acted on, and not charged, until it has been seen over time and a larger read of the source confirms it.
• A migration run shows its estimated price before you start it and is invoiced when it completes. Each cycle of a recurring sync is invoiced when it completes, without a separate quote.
• The annual platform fee is charged automatically. Its first year is charged up front, at the first billing pass after your organisation is on the Pro plan with a mapped source; when you set a source up from its connection page, or choose Pro on the billing page, the projected amount is shown first. It then renews each year on the columns mapped at that time, without a further quote. A source added during a year is first charged at the next renewal.

Prices exclude any applicable tax, which is added at invoicing according to your location.

The free tier is how the service is trialled: one connection, and a one-time allowance of mapped columns at no charge and with no card required. The allowance covers files — uploaded through the browser or read by the desktop agent — and applies only while every connection is a file; a database, warehouse or API source is billed from its first mapped column. Work beyond the allowance is billed at the standard rates above, and you will be asked for a payment method before it runs. Data migration is charged on every plan, including the free tier — the free allowance covers building the semantic layer and does not cover moving data.

Payment is handled by an external payment processor — we never see or store your card details. You can cancel at any time; cancellation takes effect at the end of the current term.

Best efforts, and your responsibility to verify

We provide the service on a best-efforts basis. The service maps and aggregates your data using an automated system, and automated mappings, translations, aggregations and query results can be wrong. Confidence scores and review queues are provided so that mappings can be checked before they are relied on.

You decide whether the service meets your needs. You are solely responsible for determining whether the service is suitable for your purposes, for verifying any output before acting on it, and for any decision you or anyone else takes on the basis of that output. We do not know your business, your obligations or the consequences of a given number being wrong; you do. Do not use an unverified output for a decision with material financial, legal, regulatory, clinical or safety consequences.

No warranties

THE SERVICE IS PROVIDED “AS IS” AND “AS AVAILABLE”, WITHOUT WARRANTIES OF ANY KIND, WHETHER EXPRESS, IMPLIED OR STATUTORY. To the fullest extent permitted by law, we disclaim all implied warranties, including merchantability, fitness for a particular purpose, title, non-infringement, and any warranty arising from course of dealing or usage of trade.

We do not warrant that the service will be uninterrupted, timely, secure or error-free; that any output will be accurate, complete or current; or that defects will be corrected. No advice or information, whether oral or written, creates any warranty not expressly stated here.

Limitation of liability

To the fullest extent permitted by law, we will not be liable for any indirect, incidental, special, consequential, exemplary or punitive damages, or for any loss of profits, revenue, goodwill, business opportunity, anticipated savings, or data — however caused and on any theory of liability — even if we have been advised of the possibility.

In particular, and without limiting the above, we are not liable for any decision made, or not made, in reliance on any output of the service, nor for any loss arising from an inaccurate mapping, aggregation, translation or query result.

Our total aggregate liability arising out of or relating to the service, under any theory, will not exceed the total fees you actually paid us in the 12 months immediately preceding the event giving rise to the claim.

Nothing in these terms excludes or limits liability that cannot be excluded or limited by law — including liability for fraud, fraudulent misrepresentation, or willful misconduct. If any part of this section is held unenforceable, it is severed and the remainder continues to apply.

Indemnification

You will indemnify, defend and hold harmless Exon Biosciences and its officers, employees and agents from any claim, demand, loss, liability, cost or expense (including reasonable legal fees) arising out of or relating to: the data you connect or upload and your right to do so; your use of the service or of any output from it; your violation of these terms or of any law; or any claim by a third party — including a data subject, regulator, employer or customer of yours — relating to data you brought into the service.

Governing law

These terms are governed by the laws of the State of Kansas, without regard to its conflict-of-laws rules. The state and federal courts located in Kansas have exclusive jurisdiction, and both parties consent to venue there.

Termination

You may delete your account at any time from Settings. If you are the last member of your organisation, doing so cancels its subscriptions, revokes the credentials we hold for its connections and deletes its data as described under deletion; otherwise only your own account is removed. We may suspend an account for non-payment or for use that endangers the service or other customers.

Changes to these terms

When these terms change, the new version and the date it takes effect are shown the next time you open the dashboard or authorise an AI assistant, and you are asked to accept them. We keep a record of each version you accepted and when. The version and effective date of the text you are reading are shown at the top of this page.

Privacy Policy

This section describes what actually happens in the system, not an aspiration.

What we collect about you

Account details you give us: name, email address, organisation name, a hashed password and, if your organisation gave them, its industry, size and region. Billing identifiers from our payment processor. Operational logs including IP address, which we use for rate limiting and abuse prevention. The questions you or your AI assistant ask through the query tools, and the queries they become, are kept in a query cache so that repeated questions are answered faster. The structure of a query — which of the platform’s shared concepts it combines and how — may be reused to answer another customer’s question. The text of your question, any value a query filters on (a supplier name, a country, an amount) and any concept that exists only in your account are never shown to another customer, and another customer’s search is never matched against the words of your question — only against the shared concepts your query uses and the kind of answer it gives (a total, a trend, a ranking). A query that filters on a value, or uses a concept of your own, is not reused outside your account.

If you sign up to try the demo workspace, we also keep the campaign tags of the link that brought you to the sign-up page, when it has any: its utm_source, utm_medium, utm_campaign, utm_content, utm_term and ref values. They tell us which of our outreach brings people to the demo; the sign-up keeps nothing else from the link. We can download a list of the people in the demo workspace other than its administrators (everyone who signed up to it, and anyone we invited into it): each person’s name, email address, when their account was created, whether their address is verified, whether and when they asked for product email, and those tags.

Email we send you

Everyone receives account email, such as the link to verify an address or to reset a password. We send product email (news, tips and information about Universal Intelligence) only to people who asked for it, by ticking the box that offers it when they signed up to try the demo. The box starts unticked. When you tick it, we record that you did, when, and the exact sentence you agreed to. Every product email tells you how to stop receiving it, and you can stop it at any time by writing to customer_service@exonbiosciences.com.

What we store — and what we do not

Query results pass through; they are not warehoused. When you, a dashboard or an AI assistant queries a connected database, warehouse, API or desktop agent, the query runs against that source and the results are returned to you, held only in a short-lived cache. For each source we keep its metadata: table and column names, the concepts your columns map to, relationships and the schema definitions we generate.

What we do keep from your data, and why. Beyond metadata, the service keeps these values from your sources, each for the purpose given:

• Samples. When a source is connected: up to ten values from each column and, for databases, warehouses and APIs, up to five example rows from each table. They are used to map the source, to check its data quality and to show you previews. When the desktop agent sends a file your plan cannot add as a connection, the file’s name, where it is on the computer the agent runs on, its column names and up to ten values from each column (each cut to 200 characters) are kept so you can add it from the Connections page. Only your admins and the person whose agent sent it can see it. It is kept until you add or dismiss it, and for at most 30 days, after which it is deleted automatically.
• Short value lists. For a column that holds a small set of repeated values, such as a status, up to thirty of those values, to describe what the column contains.
• Names matched across sources. Each distinct spelling of a name the service matches — for example every way a supplier is written in your systems — and the name it chose for that entity, so that the same company, product or person resolves to one entity.
• Details of matched entities. When the same entity appears in several sources, the value chosen for each of its details — for example a supplier’s country — with the source values it was chosen from and the ones it was not, so you can review the choice.
• Split compound values. For a column that packs several facts into one value — for example a product name that includes its supplier and purity — each distinct value and the parts extracted from it.
• Flagged rows. Rows a data-quality check flags as suspect, so you can review them.
• Uploaded files. A file whose contents are sent to us — uploaded through the browser or by the agent’s file-upload mode, or read from cloud object storage you connect as a source — is stored, and its rows are copied into a table in our database so it can be queried like any other source. (The desktop agent’s standard mode sends only column names and the samples above; the file stays on your computer.)
• Migration downloads. A file a data migration produces for you to download is kept with that migration.

We do not build benchmarks from your data. Neither your data nor the results of your queries are copied into industry benchmarks or any other shared dataset, beyond the shared pattern library described under Automated processing.

Credentials. Passwords, API keys and OAuth tokens for most connections are held in an encrypted credential vault rather than in the application database; we store a reference to them plus non-secret details such as host and username. For three kinds of connection — an AI assistant server connected as a data source, cloud object storage, and a BI tool (Power BI, Tableau, Looker or Qlik) connected to import its dashboards — the tokens, keys or secrets are stored in the application database, encrypted. Deleting a connection, or your account, revokes the credentials held in the vault and deletes those held in the application database.

Backups. The service’s databases and uploaded files are backed up every night. Each backup is kept on our server for 14 days, and, encrypted, in off-site storage for 30 days, after which the storage provider deletes it within a day. If our nightly backups stop, the three most recent copies on our server are kept until newer ones replace them, so that a good copy always exists. Deleting your account removes what is listed under deletion from the live system at once; the copies of it in backups are deleted with those backups, within 31 days (or, if our nightly backups have stopped, when newer backups replace those three copies).

Automated processing

Parts of the service use automated processing providers. Each task sends them only what it needs:

• Mapping a source — column names and types and the samples described above. This is how a column called cust_nm is recognised as a customer name.
• Matching names across sources — the distinct values of columns that name things, such as suppliers, customers or products, so that spelling variants can be recognised as the same entity.
• Splitting compound values — the distinct values of a column that packs several facts into one value, up to several thousand per column.
• Choosing a matched entity’s details — where sources disagree about one of them, such as a supplier’s country, the competing values.
• Checking data quality — a sample of rows from a table.
• Searching your data by meaning — the search text and the values being searched.
• Importing a dashboard — the contents of the file you import: its formulas, charts and the names of the fields they use. A data file packaged inside it — a spreadsheet, text or JSON file that a Tableau workbook carries, for example — contributes only its column or field names, never its values.

Your full dataset is not sent, although for a small table a sample can include every row, and name matching and splitting compound values can include every distinct value in a column. If your organisation cannot permit values to leave your environment, tell us before you connect a source — this is a real constraint for regulated data and we would rather discuss it than have you discover it here.

The shared pattern library. Some of what is learned is kept in a library shared by all customers, so the platform improves for everyone — for example that a column matches a chemical-identifier format, how an API paginates, or how a compound value splits into its parts. Many entries record only the shape of a value (for example NNNN-NN-N), but some keep examples from the source they were learned from: a sample response from an API endpoint, or a few example values of a column. It also keeps each formula from the dashboards customers import — as written, including any value written into it, such as a filter on a product name — with its translation, so that the same formula is not translated twice, and templates for exporting formulas that are built from them. The analysis of a dashboard you import, which includes its file name, is kept for your account only and is deleted with it.

Subprocessors

We engage third-party providers to deliver the service. They are described below by category and by the data they receive, rather than by naming a fixed vendor for each — providers change, and a list pinned to one supplier would be out of date the day we add or replace one.

The categories, and the data each category receives, are what we commit to. We may add, replace or remove individual providers within a category at any time, including adopting new automated-processing providers as they become available. Doing so does not change what data leaves the service or why.

Want the specific providers? We keep a current list of the named providers in each category and will give it to you on request at customer_service@exonbiosciences.com. If your organisation requires advance notice of changes within a category, ask for that in writing before connecting personal data and we will agree it with you.

Cookie Policy

Strictly necessary — always on

A session cookie keeps you signed in, and a CSRF token protects form submissions. These cannot be turned off without breaking sign-in, and no consent is required for them.

Optional — only with your consent

One service sets optional cookies, and it does not load unless you accept optional cookies:

• Website chat — the chat widget, on our website and in the signed-in service, which also reports the pages you visit to the chat service.

Site analytics — Zoho PageSense, which records which pages you visit and how you use them, including heatmaps and session recordings — is set up on our site to load only if you accept optional cookies. It is not active: our pages do not load it, and our site’s security policy does not allow it to load, so it records nothing. If we turn it on, it will run only on our public website pages, never in the signed-in service where your data is on screen; we will update this policy first and ask for your consent again.

We ask on your first visit. Declining leaves everything working; you simply will not see the chat bubble. You can change your choice at any time in Settings (Data & Privacy) or right here. Withdrawing consent is as easy as giving it: the chat stops loading, and the cookies it set on our site are cleared. If you answered our earlier cookie question, which asked only about measuring how the site is used, you are asked again, because the chat was not part of that question.

We do not use advertising cookies and we do not sell personal data.

Data processing

Where you connect data that contains personal data, you act as the controller and we act as a processor on your behalf. We process that data only to provide the service and on your instructions, which are given by the connections and settings you configure.

We keep the security measures described above, engage only the subprocessors listed above, assist you with data-subject requests so far as the service allows, and delete your data on termination as described below.

This section is a summary, not a data processing agreement. If your organisation needs a signed DPA — with standard contractual clauses, subprocessor notification terms and audit rights — contact us at customer_service@exonbiosciences.com before connecting personal data. We would rather agree terms with you first than have you discover the gap afterwards. Note that data-protection duties apply to us by law and are not affected by the limitations elsewhere in these terms.

Your rights, export and deletion

If you are in the UK, EU or another jurisdiction with equivalent law, you have the right to access, correct, export and delete your personal data, and to object to processing.

Deletion is self-serve. When the last member of an organisation deletes their account in Settings, the service cancels the organisation’s subscriptions, revokes the credentials we hold for its connections (including BI tools connected to import dashboards), and deletes its users, its connections and what we keep from its sources — samples, schemas and mappings, the names kept for matching and the query layer compiled from them, details of matched entities, split compound values, flagged rows, uploaded files and the copies of their rows, and the analyses of the dashboards it imported — together with its migrations and billing records. In the built-in dashboard viewer, everything built on the organisation’s data is deleted — its charts and the dashboards that show only its data, whether the service built them or its users built them in the viewer’s own editor, the viewer’s descriptions of its data, the viewer’s connection to that data and the roles that gave access to it — and then the accounts the viewer holds for its users are deleted; where the viewer still refers to one (for example as the creator of a chart), that account is kept with its name, username and email replaced by a placeholder, so it holds no personal data. If the viewer cannot be reached at that moment, this is retried automatically every hour until it is done, and the email addresses needed to find those accounts are kept only until then.

What deletion does not remove. Entries in the shared pattern library described under Automated processing; concepts added to the shared vocabulary; questions in the query cache, which are kept without your account attached and are not shown to anyone; a dashboard in the built-in dashboard viewer that shows none of the organisation’s data, or that also shows another organisation’s (its charts on your data are deleted); operational records that hold no values from your sources, such as audit and usage logs and the record of which version of these terms each person accepted, and of the product-email sentence each person agreed to; copies of the demo workspace’s list of people that we downloaded before you deleted your account; if you asked for product email, the copy of your name and email address held by the provider we send it with; invoices, which our payment processor keeps for accounting; the records described under Internal reporting in the subprocessor list, until they are next refreshed from the live system (each day), after which they no longer include the organisation; and backups, which are deleted as described under Backups. To have any of these removed, contact us at the address below.

For access, correction or export requests, or anything else in this document, contact customer_service@exonbiosciences.com. We aim to respond within 30 days.

Exon Biosciences · Version 2026-11-04 · Effective 4 November 2026