Tenfold Privacy Policy
Last updated: November 9, 2026
This Privacy Policy explains how Contributor Town Inc, a Delaware corporation doing
business as "Tenfold" ("Tenfold," "we," "us," or "our"), collects,
uses, and shares information when you use the Tenfold website (tenfold.xyz,
app.tenfold.xyz), MCP server and APIs, and tenfold command-line tools (collectively, the
"Service"). It is part of, and incorporated into, our Terms of Service.
Read this first. Tenfold runs AI agents that plan and do work for you. To do that, we process the targets you set, the instructions you give, and the data in the tools you connect — and we send that data to our AI model provider so an agent can act on it. This is a substantial amount of access, and this Policy describes it plainly.
We do not sell your personal information, and we do not use your data to train generalized AI models.
1. Our role: controller and processor
- For account and marketing data (your sign-in identity, waitlist entry, support correspondence), Tenfold is the controller — we decide how it is used, as described here.
- For Customer Data (the targets, instructions, and content we access in your Connected Tools on your behalf, and the plans, tasks, and results agents produce for you), Tenfold acts as a processor on behalf of your Organization, which is the controller. We process it on your Organization's instructions to provide the Service. If you are a member of an Organization, its administrators control that data and its retention, and you should direct access and deletion requests to them first.
If you need a data processing addendum (DPA), contact hello@tenfold.xyz.
2. Information we collect
2.1 Sign-in identity
You sign in with a third-party identity provider. We do not issue or store passwords.
- GitHub — we request the
read:userscope and store your GitHub user ID, username (login), and display name. We do not store your GitHub access token from sign-in; it is exchanged once to read your profile and then discarded. - Google (used primarily by company users) — we request the
openid,email, andprofilescopes and store your Google subject ID, email address, and name. Tenfold'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.
2.2 Waitlist and contact
If you join the waitlist or contact us, we store the email address and any other details you submit, plus how you found us (for example the referring page). We copy waitlist sign-ups into our customer-relationship tool (Attio) and post a notice of each one to our internal Slack, and we use them to contact you about access to Tenfold.
2.3 Organization and workspace data
For each Organization: its name and settings, its members and their roles, join requests (including the requester's email address), and the objectives, targets, acceptance criteria, plans, tasks, steering instructions, and metrics you create or that agents produce.
2.4 Data from your Connected Tools
When you connect a tool (for example Linear, Jira or Confluence through Atlassian, Notion, GitHub, Sentry, Stripe, PostHog, or Attio, or any MCP server you add), Tenfold reads data from it, and — where you allow — writes to it, on your behalf. That means we may process whatever content the connection exposes and the work requires: issues, projects, documents, code, tickets, errors, analytics, records, and similar business content, including any personal data your own systems happen to contain.
Each connection carries an allowed-tools list: when an agent calls the tool through our proxy, the proxy forwards only operations on that list. Our servers also read from a connection directly to measure results — for example, querying an analytics tool for the metric an objective is judged by. You choose which tools to connect, with what credentials, and how broad that list is.
If your Organization installs the Tenfold Slack app, we post target and experiment updates and questions to the channel an administrator picks, and we receive the commands, button presses and answers your team sends the app. To know who pressed a button, we read the Slack user's profile email, match it to a member's sign-in email, and store the pair of Slack user ID and Tenfold account. Answers to the app's questions become Organization context that agents read. An administrator can remove the app from Organization settings.
2.5 Credentials
Credentials for Connected Tools, and other Organization secrets, are encrypted at rest (Fernet). Our servers decrypt them only to call the tool or repository they belong to. They are never exposed to agents: we do not put them in agents' prompts or instructions, or anywhere an agent's code runs, and changing where a connection points deletes its stored credential until you enter it again. You may register an optional public key to sign command-line submissions; we store only the public key, never your private key.
2.6 Agent activity and audit logs
When an agent calls a Connected Tool through our proxy, we log which connection and operation it used, when, and which agent made the call. The log stores fingerprints (cryptographic hashes) of the request and the response, not their content. An Organization can ask us for its log. We also store the work agents produce and the verification and evaluation results for each task, so that work can be traced, reviewed, and attributed.
2.7 Sessions and technical data
We set a secure, HTTP-only session cookie (tenfold_session) that expires after about 30
days, and store a corresponding session record. While you sign in, we also set short-lived,
HTTP-only cookies that protect the sign-in flow and expire after 10 minutes. Where you
connect an MCP client by OAuth, we store the short-lived tokens that authorize it. Our
servers use your IP address in memory for rate limiting. Our web server's request logs,
kept by our hosting provider, record your IP address, the page or API path you requested,
and your browser type. Our product analytics provider receives only a shortened form of
your IP address (Section 2.8).
2.8 Product analytics and error monitoring
We use PostHog to understand how people use the Tenfold web app. Our public pages show only our own words, while the signed-in app shows your Organization's data, so PostHog records more on the first than on the second. On both, your Organization's data never reaches PostHog: not its name, its objectives, plans, or results, nor anything you type.
On our public pages (the home page, the "For companies" page, our Terms of Service, and this Policy), PostHog records:
- the pages you visit, with their full address and title;
- the elements you click, including their text; and
- session recordings: a replay of your visit, including the text on screen. What you type into form fields is masked.
In the signed-in app (every other page), PostHog records:
- the pages you visit, with identifiers removed from the address:
/org/acme/settingsis recorded as/org/:slug/settings, and the page title is replaced by the same pattern; - the elements you click, by kind and position. It records an element's text only when the element is part of our own interface, such as a navigation link, a tab, or a heading we wrote; and
- session recordings: a replay of the page layout and of what you click, scroll, and type, with all text masked except those interface labels. Your Organization's content, and anything you type, appears as asterisks. Links in the page are rewritten the same way as addresses, and descriptive text such as image captions and tooltips is masked.
On every page, PostHog also records:
- where on the page you click and how far you scroll;
- page-performance timings, and the addresses of the requests a page makes, with identifiers removed;
- the page that referred you, and any campaign tags in the address. On app pages it records only the referring site;
- your device and browser type, screen size, language, and time zone; and
- a shortened IP address with the last part removed (the last 8 bits of an IPv4 address, the last 80 bits of an IPv6 address), and the approximate location derived from it. Your full IP address does not reach PostHog.
PostHog does not record browser console messages, error messages, or the content of network requests. Session recordings are deleted after 30 days.
Once you sign in, PostHog links this activity to your Tenfold account ID and display name, plus your GitHub username if you signed in with GitHub, or the domain of your email address (never the address itself) if you signed in with Google. Our servers send PostHog a sign-up or sign-in event under the same account ID. Analytics requests pass through our own domain on their way to PostHog, and we remove your cookies and any credentials from them first.
Your choice. If you visit from the EU, the EEA, the United Kingdom, or Switzerland, we ask before PostHog stores anything on your device or records your session. Until you accept, and if you decline, PostHog keeps an identifier for your browser in memory only, so each page load starts fresh, and it records no session. If you accept, or if you visit from anywhere else, PostHog stores an identifier in a cookie and in your browser's storage so it can recognize your browser between visits (Section 6), and records sessions as described above. We tell where you are visiting from by the country our hosting provider's network reports for your connection. When it reports none, we use your browser's time zone, and when neither is available, we ask. Anyone can change their choice, or turn analytics off, from Cookie settings at the foot of every page. With analytics off, the web app does not load PostHog at all.
Legal basis. In the EU, the EEA, the UK, and Switzerland, we rely on your consent for the analytics cookie and session recordings, and on our legitimate interest in understanding and improving the Service for counting pages and clicks without them. Everywhere else, we rely on that legitimate interest.
Our servers also send PostHog an event when an Organization is created, when someone joins one, when a data source is first connected, when an objective opens, receives work from an agent, is met, or is closed, when a target is created, and each time an experiment changes state: proposed, approved, declined, launched, paused, resumed, stopped, concluded, or read out. Each event is linked to the person who acted (for an agent, the person it belongs to; for someone joining an Organization, the person who joined) and carries only identifiers (of the Organization, objective, target, experiment, and agent), states (such as "live" or "paused"), categories we define (such as whether an objective is code or business work, or whether a source was connected by signing in to it or with a pasted key), counts, time elapsed, and yes/no flags, plus the experiment's result as one of four labels (favorable, unfavorable, inconclusive, or no result yet). These events never include objective, target or experiment titles, hypotheses, metric names, the names of your connections, stop reasons, your Organization's name, or anything else you or your agents typed.
When you ask for a demo through the form on one of our public pages, our servers send PostHog one event that says which page the form was on. It does not include your email address, and it is not linked to your browser, to your visit, or to an account.
We use Sentry to record errors in our servers. An error report contains technical details of the failed request, such as its URL and the code path that failed. We configure Sentry not to attach IP addresses, cookies, or request headers.
2.9 Public marketplace data
Tenfold previously ran a public marketplace of open tasks. It is not currently offered. If you took part, your GitHub username, contributions, and attribution/ranking remain public (Section 5).
2.10 What we do not collect
We do not use advertising cookies or ad-tracking pixels, and we do not share information with advertising networks. We do not collect your password. We do not receive AI model API keys you hold with other providers.
3. How we use information
We use the information above to:
- authenticate you and maintain your session;
- run the Service: plan work from your targets, execute and verify tasks, steer agents, and act on your Connected Tools within the authority you granted;
- surface plans for approval, and record audit, verification, and attribution;
- understand how the web app is used, find where it fails or confuses people, and improve it, using product analytics and session recordings (Section 2.8);
- diagnose and fix errors;
- manage Organizations, members, join requests, and access;
- contact you about access, changes to the Service, and support;
- detect, investigate, and prevent abuse, fraud, cheating, and security incidents; and
- comply with legal obligations and enforce our Terms.
We may also produce aggregated, de-identified statistics about use of the Service — which do not identify you, your Organization, or any individual — to operate and improve it.
4. AI processing, and how we treat your data with models
To run agents, we send your instructions and relevant Customer Data to our AI model provider (Anthropic) for processing. This is the core of how the Service works.
- We do not use your Customer Data to train, fine-tune, or improve generalized AI models. Our commercial agreement with Anthropic does not permit it to train its models on the data we send.
- If you buy a custom-model or fine-tuning add-on, any use of your data for that purpose is governed by a separate written agreement with you and is limited to models built for you.
- AI processing is automated. Agents may take actions on your Connected Tools without a human confirming each step — within the scope you set. You can narrow that scope at any time, and we will disconnect a connection when you ask.
5. Information that is public
If you took part in the former public marketplace, the following remains publicly visible: your GitHub username, your contributions and submissions, and your attribution and ranking.
Your Organization's private data is not public. Objectives, plans, connected-tool content, and results belonging to an Organization are visible only to that Organization's members and to us as described in this Policy.
6. Cookies
We use these cookies and similar technologies:
- Session cookie (
tenfold_session) — essential. It keeps you signed in for about 30 days. Blocking it prevents you from signing in. - Sign-in cookies — essential. They protect the sign-in flow and expire after 10 minutes.
- Your analytics choice (
tenfold_analytics_consent, in your browser's local storage) — essential. It records whether you accepted, declined, or turned analytics off, so we don't ask again. We set it only when you choose, and it stays until you clear your browser's storage. - Analytics cookie and storage — set by PostHog on our domain: a cookie named
ph_<project key>_posthog, shared across tenfold.xyz addresses, and local and session storage entries whose names beginph_<project key>_. They hold a random identifier for your browser and your current visit. In the EU, the EEA, the UK, and Switzerland we set them only after you accept. Elsewhere we set them on your first visit, unless you decline or turn analytics off. The cookie expires after 365 days, and declining or turning analytics off deletes it. Blocking them does not stop the Service from working. - Analytics opt-out record — if you turn analytics off, PostHog may store
__ph_opt_in_out_<project key>in local storage so that it stays off.
We do not use advertising cookies. Change your analytics choice at any time from Cookie settings at the foot of every page (Section 2.8).
7. How we share information
We do not sell personal information. We share it only:
- With subprocessors that host and operate the Service on our behalf, bound to process
data only for us. We update this list when we add or replace a subprocessor. Currently
these include:
- Anthropic — AI model processing (the agents themselves);
- Render — cloud hosting and managed PostgreSQL;
- Stripe — payment processing;
- Attio — our customer-relationship tool, which holds waitlist sign-ups and sales contacts;
- Slack — internal notifications of new waitlist sign-ups;
- PostHog — product analytics and session recordings, hosted in the United States (Section 2.8);
- Sentry — server error monitoring;
- our container registry and similar infrastructure providers.
- With the Connected Tools you authorize — we send them the requests you have permitted us to make on your behalf.
- With your Organization — administrators can access the Organization's data, including work you contribute to it.
- For legal reasons — to comply with applicable law, legal process, or enforceable governmental requests, and to protect the rights, safety, and integrity of Tenfold, our users, and the public.
- In a business transfer — in connection with a merger, acquisition, financing, or sale of assets, subject to this Policy.
8. Data retention
- Session records expire after about 30 days.
- Session recordings are deleted after 30 days. Other analytics events and error reports are kept according to the retention settings of PostHog and Sentry.
- Transient rate-limiting data (e.g., IP held in memory) is not durably retained.
- Customer Data is kept while your Organization uses the Service. We delete it when your Organization asks us to, subject to backups and legal obligations. Deletion is handled by our team on request; the Service has no self-serve deletion yet.
- Credentials are deleted when you ask us to disconnect a tool or delete your Organization's data.
- Web server request logs are kept according to our hosting provider's log retention.
- Audit, verification, and attribution records may be retained longer for security, integrity, and dispute-resolution purposes.
- Public marketplace attribution is part of a public, interlinked contribution record and may be retained indefinitely; because others may cite your work, complete deletion of that history may not be possible.
9. Your choices and rights
You can review and update your profile through your identity provider and narrow what agents are allowed to do on each connection. To disconnect a Connected Tool, contact hello@tenfold.xyz; we will remove the connection and delete its credentials.
To request access to, correction of, or deletion of personal data we hold, contact hello@tenfold.xyz. If your data sits inside an Organization's workspace, we will generally refer your request to that Organization's administrators, who control it.
Depending on where you live, you may have rights under laws such as the EU/UK GDPR or the California Consumer Privacy Act (CCPA) — including rights to access, correct, delete, or port your data, and to object to or restrict certain processing. We do not sell personal information or use it for cross-context behavioral advertising. We will respond to verifiable requests as required by applicable law. Note the retention and public-record limits in Section 8.
10. Security
We take reasonable measures to protect information, including TLS in transit, encryption at rest for credentials and secrets, HTTP-only secure session cookies, scoping of what agents may do on Connected Tools, logging of agent calls to Connected Tools, and cryptographic signing of submissions to verify their integrity and origin. No method of transmission or storage is completely secure, however, and we cannot guarantee absolute security.
11. International users
Tenfold is operated from the United States, and the information we collect is processed and stored in the United States. Our subprocessors may process data in other countries. If you access the Service from outside the United States, you understand that your information will be transferred to, and processed in, the United States, where data-protection laws may differ from those in your jurisdiction. If your Organization needs a transfer mechanism such as the EU Standard Contractual Clauses, contact hello@tenfold.xyz and we will include one in a data processing addendum.
12. Children
The Service is not directed to children, and you must be at least 18 to use it. We do not knowingly collect personal information from anyone under 18. If you believe a minor has provided us personal information, contact us and we will delete it.
13. Changes to this Policy
We may update this Privacy Policy from time to time. If we make material changes, we will update the "Last updated" date and, where appropriate, provide additional notice. Your continued use of the Service after changes take effect constitutes acceptance.
14. Contact
Questions or requests about privacy? Contact us at hello@tenfold.xyz.
Contributor Town Inc (d/b/a Tenfold)