Privacy
Privacy Policy
Effective 2026-11-17 for new accounts · 2026-12-01 for accounts created earlier
Kismet is a developer tool. We try to collect the minimum personal data needed to operate the product, and to be specific about what that data is and where it goes. This policy explains both.
For an engineering-level explanation of how telemetry, your AI provider key, and your Blueprint code are handled, see Data & Security.
1.Who we are
Kismet is operated by Individual Entrepreneur Artem Tarasov, registration № 72930421, registered office at Davtashen 2-lane, 20/1-2, apt. 20/1-2, Yerevan 0059, Republic of Armenia (“Kismet,” “we,” “us”). Kismet is the data controller for personal data described in this policy. Contact: legal@kismetai.io.
Kismet is an independent product. It is not affiliated with, endorsed by, or sponsored by Epic Games, Inc. Unreal Engine and Blueprint are trademarks of Epic Games, Inc.
2.What this policy covers
This policy applies to:
- The Kismet plugin for Unreal Engine (the “Plugin”) when it sends telemetry, error reports, or bug submissions to our servers.
- The Kismet website at kismetai.io and the authenticated dashboard (the “Site”).
The Plugin calls your selected AI provider directly from your machine using your own credentials. That communication is covered by your provider’s privacy policy, not this one. See §8.
Optional accounts
Local use with your own provider does not require a Kismet account. Model requests use your provider connection configured in Unreal. Optional accounts support device linking, bug reports and dashboard features; they do not supply model access.
The public skills registry and publishing are disabled for this release. If you published a bundle previously, its license and downloaded copies remain unaffected. Historical registry records may retain public bundles, versions and publisher names after account deletion, with the account ownership link removed. Contact us about an existing publication.
3.Personal data we collect
3.1 Account data
When you create a Kismet account on the Site, we collect:
- Your email address (required).
- An optional display name you choose.
- If you sign in with a third-party identity provider (e.g., GitHub), the provider’s user ID and any profile fields they expose for our scopes.
- A password hash if you set a password (we never store passwords in plaintext).
3.2 Plugin installation data
When you authorise the Plugin to talk to your account, we collect and link to that account:
- A machine identifier generated locally and stored in
~/.kismet/machine_id(a random UUID — not your hostname or hardware serial). - Your operating system family (Windows, macOS, Linux), Unreal Engine version (e.g., 5.9), and Plugin version.
- A bearer token hash (SHA-256) used to authenticate the Plugin. The plaintext token exists on our servers only briefly, between the moment you claim a device code in your browser and the moment the Plugin polls for it (a window of seconds), then it is deleted.
This metadata is required for the Plugin to identify itself to your account; it cannot be opted out of without disconnecting the Plugin.
3.3 Plugin telemetry
When you link an account and enable telemetry, the current 0.10 plugin sends a restricted set of metadata. Four categories are controlled in Settings → Telemetry:
- Runs — generated run identifiers, known status values, timing, token counts and estimated usage where available.
- Tool calls — registered tool names, outcomes and duration.
- Errors — allowed error/status categories and numeric metadata.
- Crashes — allowed diagnostic metadata. Current automatic telemetry excludes free-form crash text, stack traces and editor log tails.
The current plugin filters payloads before transmission. Prompts, assistant text, tool arguments/results, asset paths, URLs and free-form errors are excluded from automatic telemetry. Device/version metadata accompanies requests. Category toggles do not remove previously stored data.
Connection activity can update device last-seen information. Local diagnostics are separate from network telemetry: the current plugin writes bounded metadata logs under Saved/Logs/KismetAI-diagnostics/. See Data & Security for controls.
Older clients may have submitted prompts, replies, tool inputs, error text, crash context or log excerpts. Historical records are not erased by upgrading the plugin. The current metadata restrictions do not describe those older records.
3.4 Bug reports
If you submit a bug report from the Plugin or the Site, we collect: the description you write, a session identifier, and any telemetry snippet you choose to attach (capped at ~1 MB). Bug reports do not respect the per-category telemetry toggles — sending a bug report is implicit consent to share the contents you wrote. Avoid pasting secrets into bug descriptions.
3.5 Contact requests
If you fill in the contact form on the Site, we collect the fields you submit (name, work email, optional company, role, team size, and use-case description), the form’s source page, and your browser user-agent string (used for spam triage). We do not store your IP address against the form record. Your IP is briefly processed by our rate-limiting layer to prevent abuse (see §7), but it is not written to the contact table.
3.6 Waitlist data
If you join the waitlist via the sign-in page, we store your email address and the timestamp of your submission. We do not log your IP address against the waitlist record. Waitlist records are deleted no later than 12 months after submission or when you create an account, whichever comes first. Legal basis: consent (the checkbox on the waitlist form).
3.7 Cookies and similar technologies
The Site uses a small number of strictly-necessary cookies and does not use third-party analytics, advertising, or tracking cookies:
- Authentication cookies (set by Supabase Auth) keep you signed in.
- kismet_landing is a session cookie that pins you to the marketing site when you click “Back to kismetai.io” from inside the dashboard.
3.8 Content you should not submit
The current plugin does not send provider credentials, project files, prompts or tool content through automatic telemetry. The provider connection still receives the context needed for your request. Bug reports and contact forms contain what you choose to submit and can include sensitive information.
Do not include credentials, private project material, special-category personal data or information about children in support submissions.
4.Current telemetry and historical records
The current plugin applies a closed metadata allowlist before automatic telemetry leaves your machine. This is separate from provider requests, which can include prompts and the code or assets needed for the task.
Older telemetry endpoints and records support richer run messages. Server-side filtering removes selected Blueprint tool-result content and known sensitive fields, but it is not a guarantee that all source content or secrets were removed. Older prompts, tool inputs, logs and crash reports can contain project information.
Support reports and attachments are explicit submissions outside the automatic metadata path. Review them before sending. Administrators can access stored submissions when necessary for support, security and service operation. Upgrading does not rewrite historical records or local files.
5.How we use personal data
We use the data described in §3 for the following purposes:
- To run the service — authenticate your account, issue and validate Plugin device tokens, deliver the dashboard, serve the auto-updater manifest.
- To compute and display your usage and cost — tokens, runs, models, and USD spend on your dashboard.
- To debug and improve Kismet — investigate errors, crashes, and failure patterns; reproduce bugs you report.
- To respond to you — reply to contact and support messages; send transactional emails (sign-in links, account-deletion notices).
- To prevent abuse — detect and stop automated sign-ups, scraping, and rate-limit violations.
- To comply with the law — respond to lawful requests, enforce our Terms, defend ourselves in disputes.
Legal bases (where GDPR/UK GDPR applies): performance of contract for service operation; legitimate interests for debugging, abuse prevention, and product improvement; consent for any future marketing communications.
6.AI training
We do not use your prompts, the assistant’s replies, or any Plugin telemetry to train, fine-tune, or evaluate AI models — ours or anyone else’s. We do not permit third parties to use this data to train their models. The only exceptions are:
- Content you explicitly report to us as feedback or as part of a bug report;
- Content flagged for security or abuse review (used to investigate the specific incident, not to train models);
- You explicitly opt in.
We may compute aggregate, de-identified statistics (e.g., “average tokens per run by model”) for our own product analytics. We do not attempt to re-identify de-identified data.
7.Sharing and recipients
We share personal data only with vendors that operate the service on our behalf, under written terms that restrict their use of the data:
- Supabase, Inc. — managed Postgres, authentication, and serverless-functions hosting. Stores all account, telemetry, and bug-report data described above. Request limits also use short-lived keyed hashes and counters in Supabase; raw IP addresses are not stored in those counters.
- Vercel Inc. — hosts the Site and processes ordinary HTTP-request metadata (IP, user agent, path) in its access logs. Vercel also collects anonymised web-vitals performance data (page load timing, route, country, anonymised IP) via Speed Insights for our own performance monitoring; no cookies are set, no cross-site identifiers are used, and the data is not sold or used for advertising.
- Discord — when operator notifications are enabled, contact details and use-case text, waitlist email and browser metadata, bug descriptions and identifiers, and error messages can be forwarded to operator channels for support and triage. Do not put secrets in these submissions. These notification copies are separate from database records.
- Cloudflare R2 — stores Plugin release archives (no personal data).
We may also disclose personal data:
- To a successor entity in connection with a merger, acquisition, financing, or sale of assets;
- To government, regulatory, or law-enforcement authorities when we are legally compelled, or to enforce our Terms or protect our rights;
- With your explicit consent.
We do not sell your personal data and we do not allow it to be used for cross-context behavioural advertising under U.S. state privacy laws.
8.AI provider relationship (BYOK)
Kismet uses a bring-your-own-key model. The Plugin calls the AI provider you select (for example Anthropic, OpenAI, or Google) directly from your machine using the API key you provide for that provider. From a data-protection standpoint:
- Your direct relationship with the provider you select governs whether and how that provider processes your prompts and code. Its policies apply, not ours, to that traffic. For Anthropic, for example, see anthropic.com/legal/privacy.
- Your API key is not transmitted to or proxied through Kismet servers and is not logged by us.
- We are not a sub-processor of your AI provider in respect of your prompts. Your AI provider is not our sub-processor for the prompt traffic either.
When the Plugin sends usage telemetry to us (separately from the call to your AI provider), the current-versus-historical telemetry distinction in §4 applies.
9.International transfers
Kismet operates from Armenia. Our hosting providers (Supabase, Vercel, Cloudflare) and notification provider Discord may process data outside your country, including in the United States and European Union. By using Kismet you understand your data may be stored or processed in those jurisdictions. Where required by EU/UK law we rely on Standard Contractual Clauses with our processors.
10.Retention and deletion
Account data, telemetry, and bug reports are retained for the life of your account. Run transcripts, events, error logs, and crash reports do not auto-expire. Current metadata-only collection does not retroactively remove older content.
Contact-form submissions are deleted no later than 12 months after submission unless they have led to an ongoing customer relationship.
Account deletion. You can request deletion at any time from Settings → Account. We mark the account for deletion and immediately revoke all Plugin tokens. Hard-deletion runs on a monthly schedule, so the actual purge happens within roughly 30 to 60 days. You can sign in and cancel while the request is waiting, until permanent deletion begins. Once deletion has begun, it cannot be cancelled. Once purged, the cascade removes your profile, machines, runs and run messages, events, error logs, crash reports, and tokens.
Database account deletion does not automatically delete notification copies already sent to Discord. Contact legal@kismetai.io to request removal of those copies. We review them separately, subject to applicable retention obligations; no automatic Discord expiry is implemented.
We may retain a minimal record of the deletion event itself (e.g., for fraud or legal-defence purposes) after the underlying account is gone.
11.Your rights
Depending on where you live, you may have rights to access, correct, port, delete, or restrict processing of your personal data, to object to processing based on legitimate interests, and to withdraw consent. You can exercise most of these from the dashboard (Settings → Account). For anything else, write to legal@kismetai.io and we will respond within 30 days.
If you are in the EU/UK and unhappy with our response, you may complain to your local data-protection authority. If you are in California, you have additional rights under the CCPA/CPRA; we do not sell personal data or share it for cross-context behavioural advertising.
12.Security
We use industry-standard measures to protect personal data, including row-level security on all user-data tables, hashed credentials and Plugin tokens (SHA-256), TLS in transit, encrypted storage at rest with our hosting providers, and least-privilege administrative access. See Data & Security for the technical details.
Because Kismet is currently operated by a sole proprietor, administrative database access is held by the operator alone and is not delegated to anyone else. The operator accesses stored data only when necessary for support, abuse investigation, security incidents, or legal compliance. We are working on per-query audit logging for that access; until it is in place we ask you to take this disclosure as the honest description of our current setup.
To report a vulnerability, write to legal@kismetai.io with subject “security”. We acknowledge reports within five business days.
13.Children
Kismet is not intended for children. We do not knowingly collect personal data from anyone under 18. If you believe a minor has provided us with personal data, write to legal@kismetai.io and we will delete it.
14.Changes to this policy
We will update this page when our practices change. The effective date at the top of the page identifies when the published policy takes effect. Material changes will be announced in the dashboard or by email at least 14 days before they take effect.
15.Contact
For privacy questions, data-subject requests, vulnerability reports, or anything else legal: legal@kismetai.io.
Postal address: Individual Entrepreneur Artem Tarasov, Davtashen 2-lane, 20/1-2, apt. 20/1-2, Yerevan 0059, Republic of Armenia.