mirrors/compiler-explorer
1
0
Fork 0
mirror of https://github.com/compiler-explorer/compiler-explorer.git synced 2025-12-27 07:04:04 -05:00
Code Issues Packages Projects Releases Wiki Activity

Update to Privacy Policy (#7983)

Browse Source 
A whole bunch of changes I've been meaning to make:
- Clarify things and put the TLDR at the top
- Remove my own darn address etc (after checking this is OK)
- Clarifying the goo.gl situation
- Paving the way for the explain feature
This commit is contained in:
Matt Godbolt
2025-08-02 16:20:20 -05:00
committed by GitHub
parent 784758f13f
commit ef264acace
3 changed files with 123 additions and 62 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
docs/Privacy.md
View File

@@ -2,7 +2,7 @@
_This is a summary of our Privacy policy, not a legal document, and might be incomplete._
_For the full Privacy policy, see `static/policies/privacy.html`, or visit https://godbolt.org/#privacy_
_For the full Privacy policy, see `static/generated/privacy.pug`, or visit https://godbolt.org/#privacy_
The main Compiler Explorer site (at https://godbolt.org/) has a cookie and privacy policy, and it's expected that any
changes to the code are in compliance with those policies. It's worth taking a look at them if you're touching any area
@@ -24,12 +24,23 @@ short links). All this makes perfect sense and would probably be done anyway, as
We anonymise IP addresses so there's no exact mapping back to an individual using an IP. Not that it's trivial to map an
IP to a user anyway.
We shouldn't store data forever: our web logs are set to delete after a few months.
We shouldn't store data forever: our web logs are set to delete after 32 days. Amazon infrastructure logs
(separate from our web logs) also contain full IP addresses and are kept for 32 days. Compilation analytics logs
are kept for up to 1 year - these contain hashed source code (not reversible), compiler options, and usage patterns
to help us improve the service. Lambda and API Gateway logs are kept for 7-14 days.
Short URLs do turn up in the web logs: from the short URL of course one can easily extract the source code embedded in
that short URL. Users are notified of this in the privacy policy. The ultimate recourse for users concerned about this
is to not use the main Compiler Explorer but instead run their own local service, which is relatively straightforward.
We also integrate with third-party services with user consent: Claude Explain sends code to Anthropic for analysis, and
we use Sentry for error reporting (which keeps IP and browser info for up to 90 days - note this is controlled by Sentry's
retention settings, not our code). Users can control whether their code is stored for diagnostic purposes through a
setting that defaults to enabled but can be disabled.
Important: The actual retention periods are configured in our terraform infrastructure, not in the application code.
Always verify that the privacy policy matches the terraform configuration when making changes.
### Admins
A very small group of people have administrator rights on the public Compiler Explorer. Those individuals can:

2
etc/scripts/parsed-pug/parsed_pug_file.js
View File

@@ -9,7 +9,7 @@ const path = require('path');
// just update the hash here.
const expectedHashes = {
cookies: '08712179739d3679',
privacy: 'c0dad1f48a56b761',
privacy: '074dd09a246ad6fe',
};
function _execGit(command) {

168
static/generated/privacy.pug
View File

@@ -17,12 +17,35 @@ html(lang="en")
| Thanks for your interest in what Compiler Explorer does with your data. Data protection is really
| important to the Compiler Explorer team, and we want to be very clear about what we do with your data.
h3 The short version
ul
li We compile your code then delete it from our servers.
li Short shared links store your code indefinitely if you choose to create them.
li We keep some logs to help run the service.
li We only share data with third parties with your explicit consent.
h3 How long we keep things
ul
li Your source code is usually deleted within minutes, or up to 1 week if we need it for debugging (unless you opt out).
li
| Short shared links are kept indefinitely. In exceptional circumstances (such as accidental exposure of sensitive information),
| we may be able to delete links. Contact
a(href="mailto:privacy@compiler-explorer.com") privacy@compiler-explorer.com
| to discuss.
li Web logs contain semi-anonymised IP addresses for up to 32 days.
li Amazon logs contain full IP addresses for up to 32 days.
li Compilation analytics (which compilers and settings are used) are kept for up to 1 year to help us improve the service.
li Error reports hold IP and browser info for up to 90 days.
li Cached compilation results are stored in memory and on disk, but can't be traced back to your original code.
h3 Who we are
p
| Compiler Explorer was created by and is primarily administrated by
|
a(href="mailto:matt@godbolt.org") Matt Godbolt
a(href="https://xania.org" target="_blank" rel="noreferrer noopener") Matt Godbolt
| , along with a number of volunteers (including, but not limited to those listed in our "
a(href="https://github.com/compiler-explorer/compiler-explorer/blob/main/AUTHORS.md"
target="_blank" rel="noreferrer noopener") Authors
@@ -35,36 +58,42 @@ html(lang="en")
|
| if you wish to help.
h3 Your data
h3 What happens when you compile code
p
| In order to process compilation and execution requests, your browser sends the source code you typed in the editor
| window along with your chosen compiler and options to the Compiler Explorer servers. There, the source code is
| written to disk and your chosen compiler is invoked on it. If your request was to have your code executed, the
| resulting executable is run. The outputs from compilation and execution are processed and sent back to your web
| browser, where they're shown. Shortly after this process completes, your source code is deleted from disk. If, in
| processing your query, an issue with Compiler Explorer is found, your code may be kept for up to a week in order to
| help debug and diagnose the problem. Only the Compiler Explorer team will have access to your code, and only for the
| purposes of debugging the site: we will never share your code with anyone.
| When you compile, your browser sends your source code and compiler settings to our servers. We write your code
| to a temporary file, run the compiler on it, and send the results back to your browser.
p
| The source code and options are also subject to a one-way hash, which is used to cache the results to speed up
| subsequent compilations of the same code. The cache is in-memory and on-disk. It's impossible to reconstruct the
| source code from the hash, but the resulting assembly code or binary output (the compilation result) is stored as
| plain text. There's no way to enumerate the in-memory cache contents. In exceptional cases, administrator members of
| the Compiler Explorer team may be able to enumerate the disk caches and retrieve the compilation output, but with no
| way to trace it back to the source code.
| Your code is deleted within minutes. As soon as compilation finishes, we clean up the temporary files.
p
| In short: your source code is stored in plaintext for the minimum time feasible to be able to process your request.
| After that, it is discarded and is inaccessible. In very rare cases your code may be kept for a little longer (at
| most a week) to help debug issues in Compiler Explorer.
| Exception for debugging: If something goes wrong with Compiler Explorer itself (not your code, but our
| system), we might keep your code for up to a week to help us fix the problem. This only happens if you have the
| "Allow my source code to be temporarily stored for diagnostic purposes" setting enabled (which it is by default).
| You can disable this in Settings if you prefer. Only the Compiler Explorer team can access this, and we'll never
| share your code with anyone else.
h4 Short links
h3 How we speed things up (caching)
p
| To make repeated compilations faster, we cache the results.
p
| We create a unique fingerprint from your code and settings using a secure hash. We cannot reconstruct
| your original code from this fingerprint. However, we do store the compilation results (the assembly or executable
| output) as plain text, linked to that fingerprint.
p
| Part of this cache lives in memory and disappears when we restart our servers. Part of it is stored on shared disk. In
| exceptional cases, the small team of trusted Compiler Explorer administrators might be able to see these cached compilation
| results, but there's no way for us to trace them back to the original source code.
h3 Shared links
p
| If you choose to share your code using the "Share" dropdown, then the user interface state including the source code
| is stored. For a "Full" link, this information is encoded into the URL as a URL hash (e.g.
| is stored. For a "Full" link, this information is encoded into the URL after the # symbol (e.g.
| #[code https://godbolt.org/#ui_state_and_code]). For short URLs, the interface state is stored on
| Compiler Explorer's servers, and a shortened name uniquely
| referring to this data is returned. The shortened name comes from a secure hash of the state, and without
@@ -73,38 +102,42 @@ html(lang="en")
| Links of this form look like #[code https://godbolt.org/z/SHORTNAME].
p
| Prior to storing data itself, Compiler Explorer used an external URL shortening service (
a(href="https://goo.gl/" target="_blank") goo.gl
| ) and the resulting short URL was rewritten as #[code https://godbolt.org/g/SHORTURLPART].
| The storage for the user experience state in this case remains with the short URL provider,
| not Compiler Explorer.
| Before September 2018, Compiler Explorer used Google's goo.gl service for short links.
| We switched to our own system in 2018, but when Google shut down goo.gl in 2025, we migrated any remaining old
| links to ensure they keep working.
h4 Application, web and error logs
h3 Application, web and error logs
p
| Compiler Explorer keeps application logs, which contain semi-anonymised IP addresses, but no other personally
| identifying information. When a long URL is clicked, the hash part of the URL is not sent to the server, so the user
| state (including the source code) is NOT exposed in the web log. If a user clicks a short URL, then the short form
| #[em is] exposed in the web log (as #[code https://godbolt.org/g/SHORTURLPART]) and from this the source code can be
| retrieved. As such, if you create a short URL of your code, your source
| code and other user state can in principle be retrieved from the web log of Compiler Explorer.
| Web access logs contain semi-anonymised IP addresses (we remove parts of the IP address to make them less identifying) but
| no other personal information.
| When you visit a long Compiler Explorer URL (the ones with #[code #] in them), your code
| stays in your browser and isn't logged. If you visit a short URL we created (like #[code godbolt.org/z/abc123]),
| then we can potentially retrieve your code from our logs.
p
| Compiler Explorer keeps a separate compile request log for Analytics purposes without identifying information.
| This log only contains the settings which were used - minus code and options that may contain sensitive data.
| Compilation logs are separate analytics logs that record which compilers and settings people use.
| These analytics help us understand usage patterns and plan improvements. We store a fingerprint (hash) of your
| source code along with compiler options, filters, and libraries used, but we can't reverse this to see your
| actual code. These analytics are kept for up to 1 year. We may share aggregate statistics about compiler usage publicly,
| but these never include individual usage patterns or any way to identify specific users.
p
| Compiler Explorer uses Amazon's web serving, load balancing and edge caching systems. In order to debug and diagnose
| Compiler Explorer, to help track down and block Denial of Service attacks, and to gather statistics about Compiler
| Explorer's performance and usage, the logs from these systems are archived. These logs contain the full IP addresses
| of requests. They are kept for no more than one month, after which they are permanently deleted.
| Amazon infrastructure logs: We use Amazon's servers to run Compiler Explorer. Their logs (which help us
| debug issues and block attacks) contain full IP addresses and are kept for up to 32 days, then permanently
| deleted.
p
| If your web browser experiences an error, we use a third party reporting system (
| For error reporting: If something goes wrong in your browser, we use
|
a(href="https://sentry.io/" target="_blank") Sentry
| ). This keeps information, including your IP address and web browser user agent, for no more than 90 days.
|
| to help us fix it. This keeps your IP address and browser information for up to 90 days.
h4 Executing your code
p
| If we need to share data with new third-party services in the future, we'll update this privacy policy accordingly.
h3 Executing your code
p
| For certain configurations, we may support executing the results of your compilation on the Compiler Explorer
@@ -112,37 +145,57 @@ html(lang="en")
| both the Compiler Explorer site and other concurrently-processed requests from information leakage due to rogue
| executions.
h4 Cookies
h3 Claude Explain
p
| The "Claude Explain" view sends your code and data to Anthropic, the makers of Claude. We always ask for consent
| before sending your code, and Anthropic does not use anything we send them for training. For the purposes of
| Anthropic's own Privacy Policy, we use them as a "data processor", and so our own privacy policy applies to the
| data.
h3 Cookies
p
| Compiler Explorer uses small pieces of information stored on your computer: Cookies and Browser Local
| Storage (and Session Storage). Local storage is used to remember the user's settings, source code and user interface configuration, so
| that it's available when the user visits the Compiler Explorer site again. This information is not transmitted to
| Compiler Explorer, except as described above in order to fulfil the user's requests. There is a
| Storage (and Session Storage). Local storage remembers your settings, source code and user interface configuration
| so it's available when you visit again. This information is not transmitted to
| Compiler Explorer, except as described above in order to fulfil your requests. There is a
|
a(href="#cookies" rel="noreferrer noopener") separate document
|
| covering more on this.
h3 Your choices
h3 Your rights and choices
p
| Compiler Explorer is an open source project. If you are concerned about any of the data protection measures outlined
| above, or about what happens to your source code, you are encouraged to run your own local instance of Compiler
| Explorer. Instructions on how to do this are on the
| You decide if and when to create shared links. In case of an emergency you can request
| deletion by contacting us.
p
| For Claude Explain, we always ask for your explicit consent before sending any code to Anthropic.
p
| Compiler Explorer is open source. If you prefer complete control over your data,
| you can run your own instance. Instructions are on our
|
a(href="https://github.com/compiler-explorer/compiler-explorer"
target="_blank" rel="noreferrer noopener") GitHub project page
| .
p
| If you have questions about your data or want to request deletion of a shared link, contact us at
|
a(href="mailto:privacy@compiler-explorer.com") privacy@compiler-explorer.com
| .
h3 Compiler Explorer and the GDPR
p
| The Compiler Explorer team believes the Compiler Explorer site is compliant with the EU's General Data Protection
| Regulation (GDPR). Specifically, we store no personally identifying information, we anonymise the little data that
| we do have and we do not permanently store any user data.
| We comply with the EU's General Data Protection Regulation (GDPR) because we don't store personal information
| long-term. IP addresses are semi-anonymised and deleted within one month, and your source code is processed
| temporarily and then deleted.
h4 Name and Address of the controller
h4 Name and contact details of the controller
p
| The Controller for the purposes of the General Data Protection Regulation (GDPR), other data protection laws
@@ -151,9 +204,6 @@ html(lang="en")
div
| Matt Godbolt
br
| 2626 Orrington Ave
| Compiler Explorer LLC
br
| Evanston IL 60201 USA
br
| +1 312 792-7931<br>
a(href="mailto:matt@godbolt.org") matt@godbolt.org
a(href="mailto:privacy@compiler-explorer.com") privacy@compiler-explorer.com