Slycer Privacy Policy

Version
21 November 2026, for Slycer 0.4.0
Address
https://slycer.ai/privacy

“We” and “Slycer” mean the company named in section 1.1. “The Software” means the Slycer program you install.

1. About this policy

1.1. Who we are

Slycer is portward LLC, a Texas limited liability company, whose business address is 5900 Balcones Drive, Ste 100, Austin, TX 78731, United States. We make the Software and run slycer.ai, the one website for it. We are the controller of the personal data this policy describes. If your firm sends us the names and work e-mails of its staff and treats us as its processor for them, write to us (section 13) and we will sign a data processing addendum for them.

1.2. What this policy covers

This policy covers the Software, slycer.ai, the sign-in and account pages at mcp.slycer.ai, licence.slycer.ai, buying and billing, the e-mails we send and receive, and support. It does not cover the agencies’ websites and services, your AI client, or the pages Stripe runs for payment. Each of those has its own policy. Section 7 links to Stripe’s, Anthropic’s and Cursor’s; each agency’s is on its own website.

Slycer sells to businesses only. Most of the personal data here is about the people who buy and use Slycer at those businesses.

1.3. This version

This version describes Slycer 0.4.0. Using it to fetch data, export files and build local copies needs no sign-in and no licence key. The paid features need you to sign in with a work e-mail address that holds a seat. The Software then keeps your plan’s licence key on your computer, checks it there and renews it, as section 2.1 says. Nobody pastes a key and we e-mail no key.

2. The Software on your computer

2.1. What the Software sends to Slycer

The Software sends Slycer only the three kinds of request below. Slycer runs no server for your requests and collects no other usage data from the Software.

  1. The list of working sources. Once a day the Software asks https://slycer.ai/sources-status.json which sources are coming soon, so that a source an agency has broken can be held back without a new release. It is one plain GET, made in the background and never as part of your request, at most once a day for all the Slycer processes on your computer. The request carries the Software’s name and version (slycer/0.4.0 in this version) and nothing else: no cookie, key, licence, e-mail address or part of your requests. Like any request, it shows your IP address to the company that hosts slycer.ai, which logs it as section 3.1 says. The reply is kept as sources-status.json in ~/.config/slycer/. To turn this off, set SLYCER_SOURCES_STATUS=0 in the environment or in the key file ~/.config/slycer/.env; the Software then makes no such request and does not read the saved copy.
  2. Signing in, only when you sign in for the paid plan. You sign in by running slycer login in a terminal, or on the sign-in page the Software opens in your browser the first time you ask for a paid feature (it does this only in a server your AI client started on your computer, and only when your computer has a graphical browser). The Software listens on your own computer, at 127.0.0.1, for your browser to come back, and opens https://mcp.slycer.ai/login in your browser. You type your work e-mail address there and open the link we e-mail to it (section 3.7); the address goes from your browser to mcp.slycer.ai, not through the Software. Your browser then returns to the Software at 127.0.0.1 with a one-time code, and the Software makes one request to https://mcp.slycer.ai/login/key with that code, a one-time check value made on your computer for this sign-in (PKCE) and the 127.0.0.1 address it listened on, and gets your plan’s licence key back. The key, code and check value are never shown or written to a log. If you never sign in, the Software makes no such request.
  3. Licence renewal, only once you are signed in for a paid plan. The key holds your plan, seats, expiry and Stripe subscription ID, and no name or e-mail address. When 10 days or fewer are left on a key from a subscription, or it ended up to 60 days ago, the Software sends that key, and nothing else, to licence.slycer.ai to get the renewed key: at most once a day in the background, once when a paid request meets a key that has ended, and whenever you run slycer licence renew. The request carries the Software’s name and version and no device ID. To turn this off, set SLYCER_LICENCE_RENEW=0; you then sign in again when the key ends. If you use the Software without signing in, it has no key and sends no renewal.

The Software never sends Slycer your requests, the series you look at, any values, the files you export, your agency keys, your BLS contact e-mail, your record of accepted agency terms, or anything you write to your AI client. The data goes from each agency to your computer. We receive any of these only if you choose to send them to us, for example in a support e-mail (section 3.5).

2.2. What the Software sends to agencies

When you ask for data, or build a local copy of a source, the Software sends requests to that agency’s servers: the series IDs, dates and filters, and the key that agency requires, if any. Every request carries the same program name, the Software’s name and version (slycer/0.4.0 in this version), and no name, e-mail address or account detail. The one exception is your BLS contact e-mail (section 2.5). Each agency also sees the IP address the request comes from, which is normally your firm’s network. Where an agency needs a key, the key links your requests to the account you registered with that agency. Slycer does not hide either one. Each agency’s own privacy policy and terms cover what it does with the requests it receives.

Some data is fetched at the time you ask as well as when you build a local copy. For example, a request for a daily U.S. Treasury par yield curve rate is fetched from the Treasury’s home.treasury.gov when you ask, and the same host serves the local Treasury copy when you build it.

The Software plans each request from the sources that hold the data, and a source can read from more than one agency’s servers. The request then goes to the hosts of the sources in the plan. In our tests (section 2.3), building the local Eurostat copy for an energy price request also read series from the European Central Bank’s data-api.ecb.europa.eu, because the Eurostat source in the Software covers ECB data too. Appendix A lists every host the Software may request.

For the International Monetary Fund (IMF), the Philadelphia Fed’s surveys, the Central Bank of the Republic of Türkiye (EVDS), JODI, the Swiss National Bank and the Swiss Federal Statistical Office, the Software sends nothing to the agency until you have accepted that agency’s terms, on a page the Software shows on your own computer or in your AI client. Section 4.2 lists the cookie that page uses.

Once you have accepted the IMF’s terms, the Software may read the IMF’s lists of indicators and countries from api.imf.org while it plans a request, to see whether the IMF has the data, even when another source answers in the end. That read is the same for every user and carries no part of your request.

2.3. Hosts we measured

On 20 November 2026 we installed the release build of Slycer 0.4.0 into an empty home folder, as a new user would have it, and recorded every network connection it made for five test requests. For each of them, the table lists every host the request reached. These five are a sample, and Appendix A lists every host the Software may contact. For two more, an IMF request and a JODI request made after accepting those agencies’ terms, we stopped every connection before it left the computer, so the test sent nothing to either agency. Those rows show the hosts the Software tried to reach, marked “attempted, blocked”.

RequestHostAgency
US unemployment rate from FREDapi.stlouisfed.orgFederal Reserve Bank of St. Louis (FRED)
Energy prices from Eurostat, with its local copydata-api.ecb.europa.euEuropean Central Bank
ec.europa.euEuropean Commission (Eurostat)
US CPI from BLS, with the local BLS copy built the way the Claude Desktop extension starts Slycerapi.bls.govU.S. Bureau of Labor Statistics
download.bls.govU.S. Bureau of Labor Statistics
www2.census.govU.S. Census Bureau
The Survey of Professional Forecasters, after accepting its termsapi.stlouisfed.orgFederal Reserve Bank of St. Louis (FRED)
www.philadelphiafed.orgFederal Reserve Bank of Philadelphia
BIS policy rates, with the local BIS copydata.bis.orgBank for International Settlements
stats.bis.orgBank for International Settlements
An IMF request, after accepting the IMF’s termsapi.imf.orgInternational Monetary Fund (attempted, blocked)
A JODI request and local copy, after accepting JODI’s and the IMF’s termswww.jodidata.orgJoint Organisations Data Initiative (attempted, blocked)

The Eurostat row’s ECB host was reached while the local Eurostat copy was built (section 2.2). The Survey of Professional Forecasters row’s FRED host was a search of FRED’s list of series while the Software planned the request, before its terms were accepted; like every request to FRED, it carried the FRED key.

With no request, the Software made no connection to any agency. An IMF or JODI request made before their terms were accepted made no connection.

For the five requests, the Software’s daily read of the list of working sources (section 2.1) was turned off, so the table shows the hosts of agencies only. In a separate check, with every connection that would leave the computer stopped, the Software, started and given no request, tried only https://slycer.ai/sources-status.json, once, and that request held no key and no e-mail address. No test signed in, so none renewed a licence key (section 2.1), and none of the five fetched Treasury rates (section 2.2). Section 2.1 and Appendix A describe every host the Software may contact.

A FRED key, a BLS key and a test BLS contact e-mail were set for every request, the way the Claude Desktop extension’s settings pass them. The FRED key went only to api.stlouisfed.org, the BLS key only to api.bls.gov, and the test e-mail only to download.bls.gov. Making a diagnostics file (section 3.5) made no connection, and the file held no key and no e-mail address.

2.4. What stays on your computer

These stay on your computer and are never sent to Slycer:

  • local copies of sources and saved answers, in the Software’s data folder (.slycer/data in your home folder unless you chose another; Terms of Service section 3.5 says how);
  • your agency keys, in one of these places, depending on where you type them:
    • with slycer setup: in ~/.config/slycer/.env, readable only by your user account;
    • in the Claude Desktop extension’s settings: encrypted, in Claude Desktop’s Claude Extensions Settings folder (the encryption key is in the Keychain on a Mac and tied to your Windows sign-in on Windows; if encryption is not available or fails, Claude Desktop saves them unencrypted);
    • in the plugin’s options in Claude Code: in Claude Code’s secure credential store (the Keychain on macOS, a credentials file only your account can read on Linux and Windows);
  • your BLS contact e-mail, in one of these places, depending on where you type it:
    • with slycer setup bls: in ~/.config/slycer/.env, readable only by your user account;
    • in the Claude Desktop extension’s settings: as plain text, in Claude Desktop’s Claude Extensions Settings folder;
    • in the plugin’s options in Claude Code: as plain text, under pluginConfigs in ~/.claude/settings.json;
  • your record of accepted agency terms, in ~/.config/slycer/terms.json with a log beside it, terms-log.jsonl, both readable only by your user account. Each entry holds the agency, the version of its terms you accepted, the address of its terms page, how and when you accepted, the name and version of your AI client, the browser’s program name and the Slycer version. It stays on your computer as your own record;
  • your plan’s licence key, once you sign in: in ~/.config/slycer/licence.key, readable only by your user account (or, where Slycer support set one up for you, in ~/.config/slycer/.env). This is the one exception to this list: renewal sends the key to licence.slycer.ai (section 2.1);
  • the last list of working sources the Software read (section 2.1), as sources-status.json in ~/.config/slycer/;
  • your saved datasets, in ~/.slycer/recipes/: the instructions to rebuild each dataset (the series, the period, the transform), never its values;
  • a record of your requests, in ~/.slycer/audit/, one file a month: the time, what was asked, the sources, agencies, series IDs, places and periods of each answer, whether it came live or from a local copy, any crosswalk used and the codes looked up, the files written and the Slycer version. It holds no data values and no keys. It is kept whether or not you have a paid plan; listing and exporting it is a paid feature;
  • a log of the Software’s recent errors (time, source, tool, error type, HTTP status and host), which a diagnostics file reads (section 3.5);
  • the files you export, in the folder you choose.

You can delete any of these at any time. Uninstalling the Software does not delete them.

We do not mark the BLS contact e-mail as secret in the extension or the plugin, because it is a contact address the Software sends to BLS with every data-file request (section 2.5). If you would rather it stayed out of your AI client’s settings, save it with slycer setup bls. The places above are where each client kept these settings when we checked in November 2026; the client decides how it stores them and may change that. We have not checked where Cowork keeps the plugin’s options, so this section does not say where keys or an e-mail typed there are stored.

2.5. Your BLS contact e-mail

The US Bureau of Labor Statistics serves its data files to programs only when the request names a contact e-mail. You give your own, either with slycer setup bls in a terminal or in the “BLS contact e-mail” setting of the Claude Desktop extension or the plugin for Claude Code and Cowork. Section 2.4 says where each of these saves it. The Software adds it only to requests to BLS’s data-file host, download.bls.gov, as part of the program name, and drops it if BLS redirects the request elsewhere. Requests to BLS’s other hosts and to every other agency do not carry it. When we measured version 0.4.0 (section 2.3), the test e-mail went to download.bls.gov and to no other host. Slycer never sends an e-mail address of its own and never receives yours from the Software.

2.6. Your AI client and tunnels

You choose the AI client that sends requests to the Software and reads what it returns, such as Claude Desktop, Claude Code, Cowork or Cursor. Your requests, and the tables and files the Software returns, pass to that client like any other tool result. Its provider handles them under its own terms and privacy policy with you: Anthropic for Claude, Anysphere for Cursor. Slycer has no access to your conversations.

ChatGPT runs on OpenAI’s servers, not on your computer, so it can reach the Software only through a tunnel: a service you set up that opens a route from the internet to the Software on your computer. If you connect the Software to ChatGPT, you run that tunnel. OpenAI’s terms and the tunnel provider’s terms cover that route. Slycer runs no relay and sees none of that traffic.

2.7. Installing the Software

How you install the Software decides what is downloaded and from where.

  • The Slycer package. For Claude Code, Cowork, Cursor and the terminal, the installer uv downloads the Slycer package from PyPI, the Python Package Index, which the Python Software Foundation runs at pypi.org and files.pythonhosted.org. uv tool install downloads it once and again when you upgrade. uvx, which Cursor and other clients use to start the Software, downloads it from the same hosts.
  • Claude Desktop. The extension file holds the Software’s code, so installing it downloads nothing more. You get the file from GitHub (section 3.3). Claude Desktop starts the extension with uv, so the last item below applies to it too. The setup command for your keys and the refresh command for local copies run in a terminal with uvx, which downloads the Slycer package from PyPI as the first item says.
  • The plugin for Claude Code and Cowork. Installing the plugin downloads it from its marketplace on GitHub (github.com). Once installed, the plugin starts the Slycer package you installed and downloads nothing.
  • Python. On every route, if your computer has no Python, uv downloads it from github.com or releases.astral.sh.

The Software needs no other Python package. PyPI and GitHub see each download and the IP address it comes from, under their own terms.

3. The website, buying and e-mail

3.1. The website

slycer.ai runs no analytics and sets no cookies. It counts no visits and loads no script that tracks you. The company that hosts slycer.ai keeps a log of each request to the site (IP address, time, page and browser), which we use only to run the site and stop abuse. Our host, Railway, keeps those logs for 7 days, its period on the Hobby plan we use (section 6).

3.2. Buying and billing

You buy through Stripe’s hosted Checkout page and manage your plan in Stripe’s customer portal. When you start the trial or a plan from our sign-in page, the Checkout page opens with the e-mail address you signed in with already filled in. Stripe collects your name, work e-mail, firm name, billing address, your tax ID if you give one (Stripe Tax uses it), and your card. We receive from Stripe your name, e-mail, firm, billing address, tax ID, your plan and its status, your invoices, the card brand, its last four digits and expiry date, and the record that you accepted the terms of service at Checkout. We never see or store the full card number.

Depending on the activity, Stripe acts for us or as a controller in its own right, such as for fraud checks, under its own privacy policy.

3.3. After you pay, and the extension download

After you pay for a plan started on our sign-in page, Stripe sends you through licence.slycer.ai back to the sign-in page at mcp.slycer.ai, which finishes your sign-in; it shows no key. The Claude Desktop extension file and the install steps in section 2.7 are described on the install pages of slycer.ai, such as https://slycer.ai/install/claude-desktop, and the extension file itself is downloaded from the latest release of Slycer’s public plugin repository on GitHub (github.com). The file holds code only, never data. The install pages ask you to sign in to nothing, set no cookie and record nothing about you beyond the request logs in section 3.1. GitHub sees the download and the IP address it comes from, under its own terms (section 2.7).

3.4. E-mail from us

Stripe sends your receipts and invoices, a reminder 7 days before your trial ends and a reminder before each renewal. When you ask for a sign-in link, we send one e-mail to the address you typed, through our sign-in e-mail provider (section 7). It holds the link, which works for 15 minutes and once, and says who asked for it (Claude, the slycer command or the account page). We e-mail no licence key. We write to you from an address at slycer.ai about your plan, changes to our terms or this policy, and problems that affect you. We send no newsletters or marketing e-mail to customers unless you ask for them.

3.5. Support

Write to us at support@slycer.ai. When you write, we keep your message, your name and e-mail address, and anything you attach, to answer you and fix the problem. Leave client names, client figures and your keys out of a message; the series ID, the countries and the period are enough to reproduce a problem.

If you choose to send a diagnostics file (slycer diagnostics in a terminal, or ask your AI client for a diagnostics report), read it before you send it. It holds the Software, Python and operating system versions, how each source answers on your computer, whether each key and the BLS contact e-mail is set (yes or no), whether each agency’s terms are accepted, the age of each local copy, and the last 50 errors with their time, source, tool, error type, HTTP status and host. It holds no keys, e-mail addresses, request text, data values, file paths or computer names. Making the file sends nothing anywhere (section 2.3).

3.6. Invitations

We may write to economic consultancies to offer a plan. For this we use a person’s name, work e-mail, firm and role, taken from the firm’s own website or other public business sources. We send none to people in Germany, where the law requires prior consent even for business e-mail of this kind. Each invitation says how to ask us not to write again, and we stop when asked. If we wrote to you, you can ask us where we got your address and have it deleted (section 9).

3.7. Signing in and seats

You sign in for the paid plan on our sign-in page at mcp.slycer.ai, with the work e-mail address on your subscription. The sign-in service (mcp.slycer.ai) and the licence service (licence.slycer.ai) receive and keep the following.

  • Your address. You type it on the sign-in page. We send it a sign-in link (section 3.4). The sign-in service asks the licence service which plan the address signs in to, and the licence service looks it up in Stripe: as the e-mail of the Stripe customer that bought the plan, or as a seat address on a subscription. If the address has no plan, the sign-in page offers Stripe Checkout with the address filled in. The sign-in service does not store your address on disk. It travels inside encrypted tokens: the link, the waiting page, and the account page’s cookie (section 4.2). After a checkout starts, the address and the sign-in it belongs to are held in memory for 2 hours so the sign-in can finish when you come back.
  • Seats. A seat is an e-mail address on the subscription. The owner (the address that bought it) adds and removes addresses on the account page at https://mcp.slycer.ai/account, up to the number of seats paid for. The licence service writes each address to the subscription at Stripe, under a key made from the first 24 characters of the address’s SHA-256 hash, and deletes it when the owner removes it. We do not e-mail an added address; the owner tells that person. Removing an address stops new sign-ins with it within about a minute; it does not end a sign-in it already made.
  • Sign-in tokens. The sign-in link works for 15 minutes and once, and only in the browser that asked for it. A sign-in from Claude lasts 30 days, then you sign in again. The tokens are encrypted and held by your browser or your AI client, not by us. To stop a used link or token from working again, the sign-in service keeps a file of SHA-256 hashes of used ones, each with its expiry time, and nothing about you.
  • Rate limits. To stop abuse, the sign-in service counts in memory, for up to an hour, the links asked for by each address and the requests from each IP address, keyed by the SHA-256 hash of the address or the IP address. A restart clears them.
  • Logs. The sign-in service writes one line per request (the method, the page and the status) and lines that say how a sign-in went, such as “signed in, by e-mail, after a checkout”. When a sign-in link carries a channel code, such as one in a post that named Slycer, or you answered the optional “How did you find Slycer?” question from its fixed list, the line that records the sign-in adds them. Once per sign-in from Claude, it writes how long it took from opening the sign-in page to the first answer, in a bucket such as “under 1 minute” or “1 to 5 minutes”. No line holds your address, name, IP address, cookie, token or key. The licence service logs what it did, such as “seat: key (monthly)” or “team: add, added”, and no address.

4. Cookies

4.1. No cookie banner

slycer.ai shows no cookie banner because it sets no cookies and runs no analytics. Of the cookies in section 4.2, one is set by the Software in your own browser, on your own computer, and two by our sign-in pages at mcp.slycer.ai, only to sign you in. Under the UK’s PECR, a cookie that is strictly necessary to provide the service you asked for needs no consent, and the EU’s ePrivacy rules have the same exemption.

4.2. The cookies

slycer.ai sets no cookies and loads no Stripe script. The Software sets one cookie in your own browser when it shows an agency’s terms page on your computer, and our sign-in pages at mcp.slycer.ai set two:

CookieSet byWherePurposeLasts
slycer_termsThe Software, on your computer127.0.0.1 (your own computer), the terms page onlyStrictly necessary: ties the agency terms page to the browser that opened it, so only your click records an acceptanceUntil you close the browser
__Host-slycer_signinOur sign-in pagemcp.slycer.aiStrictly necessary: ties the e-mailed sign-in link to the browser that asked for it, so a link forwarded or opened by a mail scanner signs no one in. It holds a random value; we keep only its SHA-256 hash, inside the linkUp to 2 hours; cleared when the sign-in finishes
__Host-slycer_accountOur account pagemcp.slycer.aiStrictly necessary: keeps you signed in to the account page. It holds, encrypted, the address you signed in with, your plan, seats and end date, and a random value that guards the page’s forms; nothing is kept on our side1 hour, or until you sign out

The slycer_terms cookie never leaves your computer. Slycer never receives it. The two mcp.slycer.ai cookies are HttpOnly, sent only over HTTPS, and not sent with requests from other sites’ forms (SameSite=Lax).

4.3. Pages run by others

Stripe’s Checkout page and customer portal run on Stripe’s own sites, and Stripe sets its own cookies there, including for fraud prevention, under Stripe’s cookie policy (https://stripe.com/legal/cookies-policy). The agencies’ websites, which you open only if you follow a link, set their own cookies under their own policies.

5. What we collect, why and on what basis

CategoryWhatWhere it comes fromWhyLawful basis
Website request logsIP address, time, page and browserThe company that hosts slycer.ai, from your visitTo run the site and stop abuseLegitimate interest
Purchases and billingName, work e-mail, firm, billing address, tax ID, plan, invoices, card brand and last four digits, terms acceptanceStripe, from Checkout and the customer portalTo sell you the plan, bill you, collect tax and keep records the law requiresContract; legal obligation
Support messagesYour messages, name, e-mail and attachments, and any diagnostics file you sendYouTo answer you and fix problemsContract; legitimate interest
Licence renewalsThe licence key the Software sends (plan, seats, expiry, Stripe subscription ID) and the IP address of the request, which is used to limit the rate of requests and is not loggedThe Software, once you are signed in for a paid plan (section 2.1)To give you a renewed keyContract
Sign-in and seatsYour work e-mail address, the e-mail addresses an owner adds to a subscription’s seats, and the IP address of each request to mcp.slycer.ai, used in memory to limit the rate of requests (section 3.7)You, on the sign-in and account pages; StripeTo sign you in, find your seat, start a checkout and let an owner manage seatsContract; legitimate interest for the rate limits
Sign-in countsLog lines that say a sign-in finished and how (e-mail, checkout, account page or key), the channel code in the sign-in link, the answer to “How did you find Slycer?” if you gave one, and how long it took from opening the sign-in page to the first answer, in buckets such as “1-5m”; none holds an address, name, IP address, token or key (section 3.7)Our sign-in serviceTo count sign-ins per channel and how long setup takes, without site analyticsLegitimate interest
InvitationsName, work e-mail, firm and rolePublic business sourcesTo offer firms a planLegitimate interest

We do not sell personal data, use it for advertising, or use anything we receive to train an AI model.

6. How long we keep it

CategoryHow long
Website request logs7 days. Our host, Railway, keeps the site’s request logs for 7 days on its Hobby plan, the plan we use (https://docs.railway.com/observability/logs).
Purchases and billingWhile you are a customer, then 7 years for invoices and tax records. Stripe keeps its own records under its policy.
Licence renewals7 days. Our host, Railway, keeps the licence server’s logs for 7 days on its Hobby plan, the plan we use (https://docs.railway.com/observability/logs). The server itself logs no client address.
Sign-in and seatsSign-in links and the waiting page: 15 minutes. A checkout started from the sign-in page, with its address: 2 hours, in memory. The account page session: 1 hour. A sign-in from Claude: 30 days, then sign in again. These live in encrypted tokens and cookies, not in a database. The used-link record on our server holds only SHA-256 hashes, each until its token would have expired. Seat addresses: on the subscription at Stripe until the owner removes them or the subscription is deleted.
Sign-in counts7 days. Our host, Railway, keeps the sign-in service’s logs for 7 days on its Hobby plan, the plan we use (https://docs.railway.com/observability/logs). The log lines hold no e-mail address, cookie, token or key.
Support messages2 years after the matter closes. Names and work e-mails of a firm’s staff that the firm sends us are deleted, or removed from the thread, within 30 days after the firm’s plan ends.
InvitationsUntil you ask us to stop, or 12 months after our last message with no reply, whichever is first. We keep only your address on a do-not-contact list after you ask us to stop.

Railway says that moving to a higher plan restores logs from before those 7 days, so Railway may hold the logs of slycer.ai, mcp.slycer.ai and licence.slycer.ai longer than it shows them to us; that is under Railway’s own policy (section 7).

7. Who we share it with

We share personal data only with the service providers that run Slycer for us, and only what each needs:

ProviderWhat it does for usIts privacy policy
StripeCheckout, billing, tax, receipts, the customer portalhttps://stripe.com/privacy
RailwayHosts slycer.ai, including the list of working sources and the install pages, and keeps the request logs in section 3.1https://railway.com/legal/privacy
RailwayRuns licence.slycer.ai, which issues and renews licence keys and finds the seat for a signed-in address in Stripe (sections 2.1 and 3.7)https://railway.com/legal/privacy
RailwayRuns mcp.slycer.ai, the sign-in and account pages (section 3.7), and keeps its logshttps://railway.com/legal/privacy
ResendSends the sign-in link e-mails (section 3.4)https://resend.com/legal/privacy-policy
Google WorkspaceCarries our e-mail at slycer.ai, including support messageshttps://cloud.google.com/terms/cloud-privacy-notice

We give personal data to a court or a public authority only when a law or a court order requires it. If Slycer is sold, we e-mail customers before their personal data moves to the buyer.

Your AI client’s provider, the agencies and the tunnel you run for ChatGPT are not our service providers. You deal with them directly (sections 2.2 and 2.6). Anthropic’s privacy policy is at https://www.anthropic.com/legal/privacy, and Cursor’s at https://cursor.com/privacy.

8. Transfers outside your country

We are based in the United States, and so are some of our service providers. If you are in the EU, the UK or Switzerland, your data is transferred to the United States. We rely on each provider’s transfer terms, such as the EU Standard Contractual Clauses or the EU-U.S. Data Privacy Framework where the provider is certified. For data from the UK, those terms are the UK Addendum to the Standard Contractual Clauses or the UK Extension to the Data Privacy Framework, and for data from Switzerland, the clauses as adapted for Swiss law or the Swiss-U.S. Data Privacy Framework.

9. Your rights

Depending on where you live, you can ask us to show you the personal data we hold about you, correct it, delete it, limit how we use it, give it to you in a portable form, or stop using it where we rely on legitimate interest. You can also complain to your data protection authority. Write to us at the address in section 13. We answer within one month of your request (GDPR Article 12(3)) and may ask you to confirm who you are first.

Some data we must keep, such as invoices, even after you ask us to delete it (section 6). If your firm bought your seat, some requests may go through your firm’s admin.

10. Security

The personal data we hold is the records section 5 lists, such as billing, support and invitation records. Our sign-in service keeps no database: what it needs travels in tokens encrypted with AES-256-GCM, and the seat list is kept at Stripe. We hold none of the series, values or files you fetch. Stripe holds card data. Only Hamza Elsiesy, the LLC’s sole member, can sign in to our accounts with Stripe, the site host, the hosts of mcp.slycer.ai and licence.slycer.ai and the e-mail providers.

If a breach puts your personal data at risk, we tell the supervisory authority without undue delay and, where feasible, within 72 hours of finding it (GDPR Article 33). If the breach is likely to put you at high risk, we also tell you without undue delay (GDPR Article 34). If the breach involves the names or work e-mails of staff that your firm sent us, we also tell your firm at its billing e-mail address without undue delay.

11. Businesses only

Slycer is sold to firms and public bodies under a business contract, not to consumers. We do not market to children or knowingly collect data about them.

12. Changes to this policy

We publish each version of this policy at https://slycer.ai/privacy with its date, and we e-mail customers at least 30 days before a new version takes effect, unless a law requires a change sooner. A release of the Software that changes what it sends, or to whom, comes with a new version published before the release. Earlier versions stay available on request.

13. How to reach us

Write to support@slycer.ai or to 5900 Balcones Drive, Ste 100, Austin, TX 78731, United States about this policy, your data or a security concern.

Appendix A. Hosts the Software may contact

The Software sends requests only to the hosts of the sources the plan for your request uses or whose local copy you build, and, once you have accepted the IMF’s terms, to api.imf.org while it plans a request (section 2.2). This is every host named in the Software’s code that it may request. Section 2.3 shows which hosts the test requests reached.

SourceHosts the Software may request
abs Australian Bureau of Statisticsdata.api.abs.gov.au
alpmed Switzerland, Austria, Portugal, Cyprusdata.snb.ch, www.pxweb.bfs.admin.ch, data.statistik.gv.at, www.ams.at, bpstat.bportugal.pt, www.ine.pt, cystatdb.cystat.gov.cy, ec.europa.eu
ameco, ecbcs European Commissioneconomy-finance.ec.europa.eu, ec.europa.eu
bcch Banco Central de Chilesi3.bcentral.cl
bea, bea_regional BEAapps.bea.gov
benelux CBS, NBB, Statbel, STATECopendata.cbs.nl, www.cbs.nl, nsidisseminate-stat.nbb.be, statbel.fgov.be, lustat.statec.lu
bis BISstats.bis.org, data.bis.org
bls BLSdownload.bls.gov (data files, with your contact e-mail, section 2.5), api.bls.gov, www2.census.gov (geography files)
boj Bank of Japanwww.stat-search.boj.or.jp
census Census ACSapi.census.gov, www2.census.gov (Gazetteer files)
destatis Destatis, Bundesbank, Federal Employment Agencywww.destatis.de, www.statistikportal.de, api.statistiken.bundesbank.de, statistik.arbeitsagentur.de
eia US Energy Information Administrationapi.eia.gov, www.eia.gov
eits Census economic indicatorsapi.census.gov, www.census.gov, www2.census.gov
estat e-Stat, Japanapi.e-stat.go.jp, www.e-stat.go.jp
eurostat, eufin, eutrade Eurostat and ECBec.europa.eu, data-api.ecb.europa.eu
evds TCMB EVDSevds3.tcmb.gov.tr
fdic FDICapi.fdic.gov, banks.data.fdic.gov
fedboard Federal Reserve Boardwww.federalreserve.gov
fred FRED and ALFREDapi.stlouisfed.org
ilostat ILOSTATrplumber.ilo.org
imf, imfx IMFapi.imf.org
inespain INE Spain and Banco de Españaservicios.ine.es, www.ine.es, app.bde.es, www.bde.es
insee INSEEbdm.insee.fr, www.insee.fr, data-api.ecb.europa.eu
ireland CSO, Central Bank of Ireland, Department of Financews.cso.ie, www.cso.ie, www.centralbank.ie, databank.finance.gov.ie
istat Istatesploradati.istat.it, www.istat.it
jodi JODIapi.publisher.jodidata.org, www.jodidata.org
mexico Banco de México and INEGIwww.banxico.org.mx, www.inegi.org.mx
nordic Nordic statistics offices and central banksapi.statbank.dk, data.ssb.no, data.norges-bank.no, statistikdatabasen.scb.se, api.riksbank.se, pxdata.stat.fi, px.hagstofa.is
oecd, oecdx OECDsdmx.oecd.org
qcew BLS QCEWdata.bls.gov
spf Philadelphia Fed SPFwww.philadelphiafed.org
statcan Statistics Canada and Bank of Canadawww150.statcan.gc.ca, www.bankofcanada.ca
treasury US Treasuryapi.fiscaldata.treasury.gov, home.treasury.gov
uk ONS and Bank of Englandwww.ons.gov.uk, api.beta.ons.gov.uk, www.bankofengland.co.uk
usda USDAapps.fas.usda.gov, esmis.nal.usda.gov
usint BEA international accounts and Treasury TICapps.bea.gov, ticdata.treasury.gov
uspop Census and UN populationapi.census.gov, www2.census.gov, population.un.org
usregional US regional and weekly indicatorsoui.doleta.gov, www.fhfa.gov, www.uscourts.gov, www.philadelphiafed.org
ustrade Census international tradeapi.census.gov, www.census.gov
wb World Bankapi.worldbank.org, databank.worldbank.org, datacatalog.worldbank.org
wbprices World Bank Pink Sheetwww.worldbank.org, data.worldbank.org

Besides these agency hosts, the Software contacts three Slycer hosts, as section 2.1 says: slycer.ai, for https://slycer.ai/sources-status.json once a day; mcp.slycer.ai, only when you sign in, for the one request that trades the sign-in code for your key (your browser opens the sign-in page there); and licence.slycer.ai, only to renew a signed-in plan’s licence key.

The code also names agency web pages, key sign-up pages and documentation. Your browser opens those only if you follow a link, and the Software never requests them.