Quire
Effective date: 3 October 2026.
Current offering and responsible parties#
Quire is operated by Nuits, a Finnish sole proprietorship operated by Bapusaheb Patil. Contact legal@quirelms.com about operator-side AI or privacy questions. The current public service is a pre-release waitlist and browser-local demo. The waitlist is not an AI-learning service. Demonstration labels, simulated results and available software adapters are not evidence of an active production AI provider or approved learner-data transfer.
In a customer deployment, the institution decides which features may be used, the permitted educational purpose and the lawful basis for learner data. A signed hosting/data-processing agreement and an actual deployment/provider schedule govern our processor role. Independently self-hosted customers configure their own endpoints and provider relationships unless separately agreed with us.
Features and information that may be processed#
Quire can support a tutor or course assistant, authoring and editing assistance, course-building tools, grading suggestions, learner insights and embeddings/retrieval. Availability depends on enabled modules, provider and model bindings, policy settings and applicable choices. These capabilities do not mean all of them are active for a particular institution.
Inputs can contain personal or confidential information:
- A tutor or assistant can process the learner's message and the course context the user is permitted to access.
- Grading assistance can process submitted work, rubric text and related assessment context.
- Learner insights can process engagement or risk indicators. Embeddings can process source material or learner messages to support retrieval.
- Authoring, editing and course-building can process selected text, surrounding context and supplied source material, including personal information an author places there.
Outputs, rendered prompts, retrieved context, provider/model identifiers, actor information and review/audit details can be recorded. A label declaring no personal data does not automatically redact arbitrary authored text. Do not enter passwords, unnecessary student records, highly sensitive health or biometric information, or material you lack authority to disclose. Where a hosted provider is used, the actual input is sent to its endpoint; it is not necessarily processed only on your device or on the institution's own server.
Notice, choices and human oversight#
An enabled feature must be identified as AI assistance and its applicable provider/use explained by the institution. Policy acceptance or a product toggle is not a substitute for a lawful basis, a provider agreement, or required child authorisation. Account → Privacy and feature-specific screens expose applicable choices; the responsible institution must explain whether a feature is optional and how to seek an alternative where required.
AI outputs can be inaccurate, fabricated, biased, incomplete or unsuitable for the learner's context. Check source material and important facts, and have authorised staff review use in teaching, assessment and institutional decisions. Do not rely on a tutor as medical, legal, safety or emergency advice. Do not assume generated text is original or free of third-party rights.
Our service policy requires meaningful human review before a consequential educational decision, including a final grade or action based on a learner-risk assessment. Grading suggestions and risk indicators assist, rather than replace, the responsible educator. We do not authorise solely automated decisions with legal or similarly significant effects without human review. A learner can ask the institution for an explanation, present their view and contest an outcome or seek review by an authorised person. An independently self-hosted customer's practices remain its responsibility and are not certified by this statement.
Teachers may read relevant course tutor conversations, and access can be recorded. A tutor conversation is not a confidential counselling channel. The institution must make that access clear to learners.
Provider approval, training and transfers#
A deployment must approve its actual provider, endpoint, model, region and contractual terms before personal data is sent. Quire's governance can check recorded provider and agreement information and policy acceptance; a configuration record is not proof that a contract has been signed or that physical processing occurs in the declared region.
We make no blanket statement that every possible AI provider never trains on inputs, retains no data, or processes only in the EU. A request parameter such as store:false does not establish every training, abuse-monitoring or retention practice. The institution and operator must evaluate the actual endpoint terms, training exclusions, retention, access and deletion assistance, and any restricted-transfer safeguards. Local or self-hosted models and customer-configured compatible endpoints have different arrangements; they are not automatically exempt from privacy obligations.
The Subprocessor Information distinguishes active appointment from available adapters. Request the deployment-specific provider information before submitting personal information. The Privacy Policy describes international-transfer requirements; this disclosure does not execute Standard Contractual Clauses or certify provider commitments.
Retention, erasure and security boundaries#
The product's tutor conversation retention is configurable and defaults to 365 days. Learner deletion of a course tutor conversation can be limited until the course ends. Several assistant, editing, insight and embedding interaction types declare a 30-day retention period, while grading-related interactions can declare 365 days. These are software defaults and feature declarations, not a universal promise for every institution, model or provider. The responsible institution must specify the actual lawful schedule, any holds and the provider's separate retention.
Removing or pseudonymising an actor identifier does not necessarily erase the rendered prompt, response, retrieved text or personal information embedded in it. Shared academic evidence, audit records and provider copies may require separate action or lawful preservation. Rights requests must assess the content and recipients, not just the account profile. Device-local demo data and downloaded copies need separate device removal.
We do not guarantee error-free filtering, automatic anonymisation, zero leakage or an independently audited AI security certification. Report unexpected disclosure or an inappropriate result to the institution and, for the part operated by us, legal@quirelms.com. Include the feature and relevant time or reference, but avoid copying other people's sensitive inputs into an ordinary email.
Children and rights requests#
Public waitlist/demo use is adult-only. Under-13 institutional use requires approved onboarding, verified school/parent authority and a signed child-data agreement. AI remains disabled for that deployment until specifically assessed and approved; a product setting alone is not authorisation. See the Age and Children's Data Policy.
For institutional AI records and educational review, contact its privacy officer or use its available Account → Privacy request workflow. For our operator-held data, contact legal@quirelms.com. We assist institutional controllers under their instructions and provide operator-side rights handling as described in the Privacy Policy, including correction, objection, applicable erasure/portability, consent withdrawal and complaint options.