Privacy Policy

This policy describes how KaldenFrame Labs handles information associated with this website, communications, purchases, and connected services.

Last updated September 29, 2026

Information we process

Depending on how you interact with KaldenFrame Labs, we may process:

basic website request and operational data, such as IP address, browser or device information, timestamps, requested URLs, and aggregate traffic measurements processed through our hosting and analytics infrastructure;
information you provide when you email us or otherwise contact us, including your contact information and message contents;
order, customer, transaction, and support information associated with purchases. Payment credentials are handled by the payment provider presented at checkout; KaldenFrame Labs does not need to receive full payment-card numbers;
information from services you authorize KaldenFrame Labs or KaldenFrame WorkHorse to access through an API or OAuth connection.

Contact brief builder

The Contact page can prepare plain-text email content and a mailto: link in your browser. Field values are not uploaded to KaldenFrame Labs, stored by the website, or included in analytics. If you choose to open the email draft, the prepared text is handed to your email client; it is transmitted to KaldenFrame Labs only if you then send the message. Ordinary page-request data and email-provider processing remain subject to the other terms in this policy.

Paid-offer buttons use one fixed /request/…/ path per public offer. That path may be counted as an ordinary aggregate page request, then redirects to the Contact page with the public route and offer identity in a browser-only fragment. URL fragments are not sent in the HTTP request. No typed inquiry field, dataset name, schema, file path, search text, customer material, or other free-text value is added to that route.

Product catalog search

The Products page matches and filters its fixed offer catalog entirely in your browser. Typing or changing a filter updates the current URL through browser history without making a new request. The query and filter are therefore available for sharing and restoration; if you load, reload, or share that URL, those values become part of the requested URL and may be processed as ordinary website request data under this policy. They are not separately submitted to a KaldenFrame application, placed in browser storage, or included as an analytics field by the catalog script.

Global site search

The Search page ranks a curated static index entirely in your browser. Search text is not submitted to a KaldenFrame application server, placed in the page URL, persisted in browser storage, or included as an analytics field. Loading the Search page and its JavaScript still creates ordinary page and asset requests through the hosting and analytics infrastructure described above. Clearing, refreshing, navigating away, or closing the page removes the current query.

Release Boundary Review fit check

The fit checker classifies the boundary answers you select and prepares an optional plain-text intake brief entirely in your browser. It does not fetch or inspect the repository URL, transmit field values, persist answers, or include them in analytics. Opening the generated mailto: link hands the reviewed brief to your email client; KaldenFrame Labs receives it only if you deliberately send that message.

Release Boundary Review intake

A request for the Release Boundary Review may include your contact information, a public GitHub repository URL, the named release workflow or path, the decision you need the review to address, and non-sensitive constraints you choose to provide. KaldenFrame Labs uses that information to assess fit, confirm scope, arrange payment, perform the accepted review, deliver the report, and support the included follow-up.

The standard service does not require private repository access, credentials, secrets, private source code, customer data, or regulated data. Do not send those materials. The public repository and related public records remain hosted by their respective providers; KaldenFrame Labs may retain the accepted intake, working notes, delivered report, and transaction records as reasonably necessary for delivery, support, accounting, security, and dispute handling.

Evidence Preflight

Evidence Preflight runs in your browser. Text entered into the worksheet is not uploaded to KaldenFrame Labs, sent to an application server, or stored by the utility. Markdown and JSON exports are generated locally on your device. Ordinary website request data is still processed when the page and its assets are requested, as described above. Refreshing or closing the page clears worksheet text unless you export it first.

Artifact Admission

The Artifact Admission package runs in your local Node.js environment and performs no network requests. It reads only the policy and workspace paths supplied by the caller and writes a decision only when an output path is requested. Downloading the package, source, or hosted schemas still produces ordinary website or GitHub request data governed by the relevant service as described above.

Attempt Boundary

The Attempt Boundary package runs in your local Node.js environment and performs no network requests. It launches only the commands named in the plan, reads and hashes explicitly protected local paths, and records command-output byte counts and SHA-256 digests without storing the output content in its receipt. Plans are executable configuration and remain under your control. Downloading the package, source, or hosted schemas produces ordinary website request data as described above.

Local Input Receipt

The preserved private Local Input Receipt implementation is no longer offered for new standalone orders. When run in a local Python environment, it makes no runtime network request, upload, telemetry call, plugin load, or credential request. It reads one explicitly selected local regular file or directory, rejects symbolic links, hashes regular-file bytes, and records only a source display name, portable relative paths, byte lengths, SHA-256 identities, user-declared limits, and explicit non-claims. It does not record absolute source paths, usernames, host identifiers, filesystem timestamps, or file contents. Those local records remain on the operator's machine unless the operator chooses to move them.

Tabular Shape Receipt

The withdrawn private Tabular Shape Receipt implementation runs in a local Python environment and makes no runtime network request, upload, telemetry call, plugin load, or credential request. It reads one explicitly selected local UTF-8 CSV, JSONL, or NDJSON file under declared parsing and resource limits. The local receipt retains source basename, byte length, exact SHA-256, parsing rules, row and column counts, ordered top-level column names, value-category counts, blank, null, missing, duplicate, row-shape-error, and bounded positive timestamp-interval observations. It does not write raw rows, samples, extrema, quantiles, frequent values, individual row hashes, or absolute source paths. Those local records remain on the operator's machine unless the operator chooses to move them.

No new standalone orders or purchase intake are accepted. Do not send datasets, schemas, column names, paths, files, credentials, private code, customer or regulated data, or private strategy and research material when asking about the retained record.

Replay Input Preflight

The licensed Replay Input Preflight runs in your local Python environment and performs no replay execution, runtime network request, upload, telemetry, plugin loading, or credential handling. It reads only the contained paths declared by the local contract and records relative paths, byte counts, SHA-256 identities, discovery counts, statuses, and issue details in the local output pack. Those records remain on the operator's machine unless the operator chooses to move them.

Purchase intake requires only the purchasing legal entity, billing contact, country, intended Python version, and approximate declared file count plus largest file size. Do not send replay inputs, datasets, credentials, private code, customer or regulated data, or private strategy and research material.

The licensed Target-Blind Feature Freezer runs in your local Python environment and performs no runtime network requests, uploads, telemetry, plugins, or credential handling. Its interface accepts a strict policy and values-free column schema, not dataset rows or target values. The local output intentionally contains declared column names, roles, normalized types, selection and exclusion reasons, and SHA-256 identities. These remain on the operator's machine unless the operator chooses to move them.

The browser-local Feature Freezer fit checker accepts only eight enumerated labels. It has no schema, dataset, path, code, feature-name, identity, account, or free-text field; performs no lookup, upload, or submission; and does not persist or transmit the selected answers unless the visitor deliberately sends the generated email brief.

The purchase request asks only for non-sensitive billing and runtime-fit information. Do not send datasets, target values, confidential feature names, credentials, private source code, customer information, regulated data, or material you are not authorized to disclose.

Temporal Boundary Auditor

The licensed Temporal Boundary Auditor runs in your local Python environment and performs no runtime network requests, uploads, telemetry, or credential handling. It reads only the contained relative CSV, JSON Lines, and contract paths selected by the operator. Its reports contain declared column names, aggregate counts, rule identifiers, byte counts, and SHA-256 values, but do not copy row values, entity values, row timestamps, or absolute source paths.

A purchase request may include the purchasing legal entity, billing contact, country, intended Python version, and required CSV or JSON Lines input formats. Do not send datasets, credentials, private source code, customer data, or regulated data. KaldenFrame Labs uses purchase, delivery, and support information to assess fit, arrange payment, deliver the licensed archive, provide the included support round, keep accounting and transaction records, and address security or disputes. Public synthetic examples contain no customer data.

Temporal Boundary Auditor fit check

The product fit checker accepts only eight enumerated compatibility selections. It has no dataset, schema, path, code, identity, account, or free-text field; performs no network lookup; and does not persist or transmit the selected answers. It prepares an optional plain-text brief from those enumerated labels in your browser. Opening the generated mailto: link hands that brief to your email client; KaldenFrame Labs receives it only if you deliberately send the message.

Google and YouTube data

KaldenFrame WorkHorse uses Google OAuth and YouTube APIs to operate the KaldenFrame Labs YouTube presence. Depending on the permissions granted, this may include access to channel information, videos, playlists, ratings, comments, captions, and YouTube Analytics information. We use that information only to publish, maintain, moderate, and analyze KaldenFrame Labs' own YouTube presence and related operations.

Google user data obtained through Google APIs is not sold and is not used to build advertising profiles. OAuth credentials and machine credentials are kept outside the public website and public source repositories. Google access may be revoked through the associated Google Account. KaldenFrame WorkHorse's use of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements.

How we use information

We use information to operate and secure the website and services, provide and improve products, fulfill purchases, respond to communications, maintain customer support, measure technical and commercial performance, prevent abuse, and comply with applicable obligations.

Service providers and disclosure

We may use service providers for infrastructure, email, payments, analytics, identity, search and discovery, and other operational functions. These providers process information as necessary to provide their services. We may also disclose information when reasonably necessary to comply with law, protect rights or security, or address fraud or abuse. We do not sell personal information.

Retention and security

Information is retained only as reasonably necessary for the purposes described above and for operational, security, accounting, dispute, or legal requirements. We use technical and organizational safeguards appropriate to the information and service involved, but no system can guarantee absolute security.

Your choices

You may contact us regarding personal information associated with your interactions with KaldenFrame Labs. Where a connected service provides its own authorization controls, you may also revoke access through that service. Certain records may need to be retained for legal, accounting, fraud-prevention, or security reasons.

Changes

We may update this policy as KaldenFrame Labs' products, integrations, or data practices change. The current version will be posted at this URL with an updated date.

Security boundary

The Security & Data Handling page maps each public, browser-local, local-runtime, review, and email surface to what it reads, transmits, retains, and omits. It also states the deployed HTTP controls, vulnerability-reporting route, hash boundary, and non-certification claims.

Contact

Choose a contact route or email support@kaldenframelabs.com.