privacy
What happens to a trace you upload.
Last updated 2026-11-11
This page describes the service as it is built today, not as it is planned. There are two ways to use Boxdawn and they are not the same: a one-off upload, where you are not signed in and nothing about your trace is kept, and a connected project, where you are signed in and we keep measurements over time. Signing out is how you choose the first one. Where a section below applies to only one of them, it says so.
On this page
- 01Optional account, no tracking cookies
- 02What happens to the file itself
- 03Where the analysis runs
- 04Where your account and measurements are kept
- 05Counting uploads for the free limit
- 06What could reach a model provider
- 07The launch notice list
- 08Spend alert emails
- 09The guard plugin for Claude Code
- 10If you give the guard an API key
- 11The guard's block record
- 12Trace bodies are never in the report
- 13What the report does contain
- 14If the analysis fails
- 15Limits
- 16Questions
01Optional account, no tracking cookies
You can use Boxdawn without signing in, and that is the default. Signing in is optional, and it sets cookies whose names begin with sb-, which carry your session and nothing else. To see how people find the site, we count page visits with Vercel Web Analytics, run by our host. It sets no cookie. For each visit it records the page, the site you came from, your area (city and country), browser, device type and operating system, and the code that tells visits apart changes every day, so it does not follow you to the next day or to other sites. Your IP address is not kept in the analytics data. Everything after "?" in the address is removed before it is sent. No other analytics tool or third-party tracker is loaded, so no other cookie is set. Our hosting providers keep their own standard request logs, as every web host does.
02What happens to the file itself
An uploaded file is sent over HTTPS to the analyzer, written into a temporary directory, analyzed, and that directory is destroyed when the request ends. We do not copy it anywhere of our own, and we do not use it to train anything. On a one-off upload that is the whole story for the file: the run happens inside that single request, and nothing about the file outlives it. What is kept about the upload itself, a count for the free limit, is in the section after next.
A connected project is different. The run is handed to Modal as a background job so this page can poll for the result, and Modal keeps that job's result fetchable for up to seven days before it expires.
The boxdawn submit command sends the file to that same hosted analyzer with a project key, so it takes the same background path. From boxdawn 0.5.21 it also sends the sub-agent trace files Claude Code wrote for that session, in the same request, and they are handled exactly as the session's own file is.
On that path the file itself travels to Modal as the arguments of that job, and Modal's client puts arguments above a size threshold into its own blob storage rather than passing them inline. In the client we read, that is 2 MiB for any call at all and 8 KiB for a background call like ours. The platform can raise either, and for our own function it does not: we asked the service on 16 September 2026 and it sends no such override, so those two figures are what apply. How long Modal keeps that copy we do not know. The seven days above is documented for the job's result, we have found no published figure for the arguments, and we would rather write that down than guess.
03Where the analysis runs
The hosted analyzer runs on Modal, a third-party serverless platform, so an uploaded trace leaves your machine. Modal routes uploads through its servers in Virginia, USA, which is its default region. We have not fixed the region the analysis itself runs in, so we do not promise a country here. Transfers out of the EU are covered by the standard contractual clauses in Modal's data processing addendum, which comes with its terms rather than being something we negotiated.
04Where your account and measurements are kept
Accounts and measurements are stored in Supabase, in Seoul, and that one we do fix. The analysis happens elsewhere, and neither the trace nor the report is stored there; what is kept afterwards is kept in Korea. Korea is one of the countries the European Commission has recognised as providing adequate data protection, in a decision adopted on 17 December 2021, so personal data can move there from the EU on the same footing as a transfer inside it.
05Counting uploads for the free limit
The hosted analyzer limits free use, so it counts uploads. An upload made without signing in, or signed in with no project selected, is counted per connection: we store a keyed hash (HMAC-SHA256, with a secret key held on our server) of the IP address the request comes from, the date in UTC, and how many uploads that connection made that day. Nothing else is in that record. The address itself is not stored or written to our logs, but the same address always gives the same hash, so we treat the hash as personal data. A daily job deletes each day's counts, so none is kept for more than three days.
An upload to a project, from this page while signed in or with an API key, is counted per organization instead: we store the organization's ID, the month and the count, nothing about the person, the key or the file, and keep it for 13 months. An upload is counted once it passes the size and format checks and before it is analyzed, so a run that fails after that still counts. The public sample trace is never counted. These counts are kept with accounts, in Supabase.
Modal, which runs the analyzer, receives the IP address of every request, as any server does; whether its own request logs keep it, we have not checked.
06What could reach a model provider
The deterministic detectors do the analysis on our own machines. The semantic-duplicate LLM-judge is never enabled on the hosted analyzer, on any plan.
One further layer is different: a verification check that asks a model whether a session confirmed what it changed. Our own API key is present on the hosted analyzer as of 3 September 2026, and on a paid project this check runs on every upload, automatically. It is on by default, and a project admin can switch it off for that project; while it is off this check never starts, so this check sends nothing from that project's traces to a model, and waste measurement keeps working as before. There is no request parameter that turns it off for a single upload, and a free project is refused before the call is made.
The command line sends files with boxdawn submit. It sends the file to the same hosted analyzer with your project key, so the hosted rule above is the one that applies: on a paid project with the check on, it runs there, on our key, whether or not you asked for it.
When the hosted check runs it does not send a summary: tool names go up with their inputs, which include file paths and commands exactly as they appear in your trace, plus up to 2,000 characters of each tool's output, inside a rendered view capped at 120,000 characters. Only the main session goes into that view: from boxdawn 0.5.21 submit also uploads a session's sub-agent files, which are analyzed for cost and resends and are not sent to the model. It also carries what you asked for: the text you typed to the agent, de-duplicated and in the order it appeared, at the top of the view. Tool results are not included, even though the trace format records them under the user role. The model's answer is not written to our database, so this is a transfer and not a second copy. That is the same data we keep out of our database on purpose, which is why this section names it instead of summarising it.
Copies of boxdawn 0.5.23 and earlier also include boxdawn analyze, which runs on your own machine and sends nothing to a model unless you pass --verification yourself, with your own key.
07The launch notice list
If you leave an email address on the launch page, we store three things: the address, the language of the page you were on (English or Korean), and the time we received it. We do not store your IP address or browser details in that list. It is kept in the same database as accounts.
We use the address for one purpose: to email you when Boxdawn launches. There is no automatic deletion and no fixed retention period yet. To have your address deleted, write to contact@boxdawn.com and we delete it by hand.
08Spend alert emails
An owner or admin of your organization can set a spend alert for a project. We then store, with the project and in the same database as your account: whether the alert is on, its daily and monthly limits, and the chosen member's user ID and sign-in email. The recipient is picked from your organization's members; the site never asks for a typed address.
Every 15 minutes we compare the project's analyzed cost with the limits. The first time a day or a month passes a limit, we send one email from no-reply@boxdawn.com through Resend, an email delivery service. Resend receives the recipient's address, the subject, and the body: the start of the period, the amount, the limit and the number of runs, plus a key that stops the same email being sent twice. The project's name is not in the email. If the recipient leaves the organization, their address and user ID are removed from the alert, and it only records until someone else is chosen.
09The guard plugin for Claude Code
The Boxdawn guard is a Claude Code plugin that checks a command before it runs. It comes in two editions. The edition in our own plugin marketplace, from the boxdawn/guard repository on GitHub, carries the guard program inside the plugin, so installing and updating it are downloads Claude Code makes from GitHub. The edition for Anthropic's plugin directory, from the boxdawn/guard-plugin repository, does not.
Instead, when a Claude Code session starts and the guard program for its version is not yet in the plugin's data folder, it downloads one file for your machine from the boxdawn/guard release on GitHub, checks its SHA-256 fingerprint against the value written in the plugin, and uses it only if they match, then deletes the downloaded file. A file that does not match is discarded. This download happens once per version, with or without an API key. The request to GitHub carries no command, file path or key; as with any request over the internet, GitHub receives the IP address it comes from. Until the download succeeds, the guard cannot run, and Claude Code blocks Bash and PowerShell commands.
This section describes version 0.4.7 of the directory edition, which is in review for Anthropic's plugin directory, and version 0.4.10 of the marketplace edition. The check runs on your own machine. Both editions also check the file path of a Read or Grep call, on your own machine. Apart from that download, without a Boxdawn API key, which is the default, the guard makes no network connection while you use it, and the commands it checks do not leave your machine.
The GitHub Copilot CLI edition (0.4.10) sends nothing to Boxdawn and keeps no block record. Copilot CLI installs and updates it from GitHub, and the commands and paths it checks do not leave your machine.
10If you give the guard an API key
If you enter a Boxdawn API key in the plugin's settings, the guard makes one request in the background when a Claude Code session starts. Its purpose is to confirm that the key belongs to a project on a paid plan and, if it does, to send the plugin its current rules.
What is sent: the key, in the Authorization header, and an empty body. The command you were about to run, file paths, your conversation and any record of what the guard decided are not sent with it; the block record, described in the next section, goes in a separate request. As with any request over the internet, the service also receives the IP address the request comes from; our code does not record it.
What we keep: our server checks the key by its hash, looks up the project and plan, and records the time the key was last used. That time is stored on the key's record, in the same database as your account (see where your account and measurements are kept), and any use of the key updates it, not only the guard. Revoking a key stops the updates but keeps the record, including that time. A revoked key can be deleted from your key list, which removes its record, including the last-used time, for good. You can also email contact@boxdawn.com from the address of the account the key belongs to, and we delete that key's record within seven days.
A revoked key no longer works, and the next time a Claude Code session starts, the guard is refused and deletes its rules file. The site does not yet offer a way to delete a project or an account. Measurements are kept with the project, not with the key, so none are lost. Our code keeps nothing else from this request, except that if the lookup fails the server writes one line naming the kind of error, without the key, the command or any path, to its log. The request goes to the same Modal-hosted service as an upload (see where the analysis runs), and whether Modal's own request logs keep it, and for how long, we have not confirmed.
What is kept on your machine: if the project is not on a paid plan, the request is refused and nothing is returned. If it is, the server returns a signed rules bundle, which the plugin saves as rules.json in its data folder. The bundle does not contain your key, only a fingerprint derived from it (an HMAC), which lets us tell which key a bundle was issued to if a copy turns up elsewhere. After seven days the bundle expires: the guard stops using it and falls back to its built-in rules, but does not delete the file. The plugin deletes the file when our server refuses the key. If you remove the key from the plugin's settings, the plugin stops asking and leaves the file, so it keeps using those rules until they expire.
Team rules: when an owner or admin of your organization adds a rule on the Team rules screen, we store it with your organization, in the same database as your account: whether it stops or asks, the command name, its arguments and flags, the note they wrote, who added it and when. The signed rules bundle carries your organization's rules to the guard. A rule is turned off rather than deleted, so that block records naming it can still be read.
11The guard's block record
While the guard decides with the signed rules from a paid plan, it keeps a record of what it stopped: each time it blocks a command or asks you about one, it writes one line to a dated file in a log folder inside its data folder. A line records the tool (Bash or PowerShell), whether it blocked or asked, whether an ask was turned into a block, the IDs of the rules that matched, what kind of target was involved (for example inside or outside the project), the time, when the rules expire and the plugin build. There is no field for the command itself, file paths, file names, your user or computer name, or the session. Commands it lets through are not recorded, and on the built-in rules, with no key or an expired bundle, nothing is recorded and nothing is sent. Read and Grep decisions are not recorded.
Up to version 0.1.1, the record is not sent anywhere. From version 0.2.0 of the marketplace edition, and in version 0.4.7 of the directory edition now in review, it is sent to Boxdawn by default, for a team view of what was stopped. It is sent only when a Claude Code session starts, and only right after our server has accepted the key and returned the signed rules. What is sent: the key, in the Authorization header, the lines exactly as described above, and one random batch ID per send, which we use only so that a batch sent twice is stored once. A line that does not have exactly that shape is dropped by the plugin before sending, and again by our server. Once our server confirms it has received a batch, the plugin deletes those lines from your machine. If the server cannot be reached or reports an error, the lines stay on your machine and the plugin tries again when the next session starts.
What we keep: each line, which API key it came with (the key's ID in our database) and when it arrived, stored with your project in the same database as your account (see where your account and measurements are kept). As with any request over the internet, the service also receives the IP address the request comes from; we do not store it. Lines are kept for the detail retention period set on your project, counted from the time the command was blocked, and a daily job deletes older ones. If a key is deleted, the lines it sent stay with the project, carrying an ID that no longer points to any key; they are deleted with the project. Members of the project can see these lines on the Guard stops screen in the app. As with the rules request, if storing fails, the server writes one line naming the kind of error to its log, and whether Modal's own request logs keep the request we have not confirmed.
To keep the record on your machine only, turn on keep_blocks_local in the plugin's settings: nothing is sent, and the record stays on your machine as it does in 0.1.1. On your machine, record files older than 30 days, including lines still waiting to be sent, are deleted the next time the guard checks a command, and the whole log folder is deleted when our server refuses the key. Claude Code's own help text says that uninstalling a plugin removes its data folder unless you pass --keep-data; we have not tested this.
12Trace bodies are never in the report
The analyzer is invoked with --no-snippets, so prompt text and tool output from your trace are not included in the report at all.
13What the report does contain
A report can contain file paths and commands taken from your own trace: which file was read twice, which command ran again. Those are not masked, because that is the finding.
The report is returned to the browser that asked for it. On a one-off upload that is the only place it goes: no share links, no history, nothing to come back to.
On a connected project we do not store the report file either, but it does not disappear when you close the page: the report comes back as the result of a background job on Modal, and Modal keeps that result fetchable for up to seven days before it expires. The ticket your browser uses to fetch it does not expire on its own, so anyone holding that ticket can fetch the report during those seven days. What we keep is a set of measurements derived from it (counts, sizes, costs) plus salted hashes of the targets.
14If the analysis fails
A failure returns diagnostic output. That output is sanitized before it is returned or written to logs: caller-side file paths, email addresses, and secret-shaped strings such as API keys and bearer tokens are masked. The report body itself is deliberately left unmasked, for the reason above.
15Limits
One file per analysis, .json or .jsonl. Up to 10 MB on a one-off upload, which runs while you wait; up to 16 MB on a connected project, where the run happens in the background and this page polls for the result. From boxdawn 0.5.21, boxdawn submit sends a Claude Code session together with its sub-agent files, and the 16 MB counts them all together.
16Questions
Email contact@boxdawn.com, or open a GitHub issue.
Claims on this page describe implemented behavior. We do not claim encryption at rest, anonymization, or regulatory compliance that we have not built and verified.