Privacy policy

Last updated: October 9, 2026.

This policy covers the open-source ajq command-line tool, the ajq coding-agent skill, and this documentation website.

Data collection

ajq does not send usage telemetry or input data to its maintainer. The coding-agent skill has no hosted service and does not collect information for the maintainer. Installing the skill does not install the ajq executable, provision a model, or authorize a cloud backend.

The CLI processes the JSON or NDJSON that you provide. Those records may contain personal information if you include it. ajq uses the records to run your query and produce its output; it does not sell that information or use it for advertising.

Local and cloud processing

Ordinary jq queries do not call a model backend. The mock backend also works without network access or a model.

Semantic operations send the values being judged and their operation instructions to the backend you select. The managed local backend runs on your computer. Ollama uses the server you configure, which may be local or remote. If you select a cloud backend, such as OpenAI, OpenRouter, or Anthropic, the selected provider receives the semantic request. Its own privacy policy, retention rules, and your account settings govern that processing. Selecting a custom endpoint sends the request to that endpoint’s operator.

Model or engine provisioning and software installation may contact the download hosts used by those commands. Those hosts receive normal download-request information, such as your IP address; they do not receive your query input as part of a download.

Local storage and retention

By default, ajq stores successful semantic judgments in a cache on your computer. Cache entries include the judged values, operation instructions, model identity, results, and creation time. They can contain personal information from your input. The cache has no automatic expiration and remains until you clear it or remove its files.

Use --no-cache to bypass cache reads and writes for a run. Use ajq cache status to locate the cache and ajq cache clear to remove cached judgments. Clearing the cache does not remove your input files, saved output, installed engines, or models. See Manage the cache for these controls.

Input files and any output you save remain under your control. ajq does not delete them after processing. Your shell, coding-agent environment, configured server, or chosen model provider may keep their own records under their own policies.

Documentation and support

The documentation website does not configure visitor analytics or advertising tracking. GitHub Pages hosts it; GitHub may process normal web-request information under GitHub’s privacy statement. The site also loads fonts from Google Fonts. Your browser sends normal request information, including your IP address, to Google when fetching those fonts; see Google Fonts’ FAQ.

If you open a public GitHub issue or discussion, the information you submit is visible to the maintainer and others. Do not include private input data, credentials, or sensitive cache files. GitHub’s own policies govern the storage of those posts; ajq does not set a separate retention period for them.

Contact and changes

For questions about ajq’s data handling, use the project issue tracker. Changes to this policy will be published on this page with an updated date.