Bloombot — AI course assistant

Terms & conditions

The agreement between you and us for using Bloombot.

Last updated 18 September 2026

These terms are an agreement between you and Bloombot ("we", "us") covering your use of Bloombot ("the service"). By creating an account, joining a course, or asking the service a question you accept them. If you are using it on behalf of an institution, you confirm you may bind that institution.

What this service is

Bloombot answers students' questions about a course, on Discord and on the web, using material an instructor supplies. It is actively developed software run as a teaching pilot, not a finished product. Read the availability section below before you depend on it for a class.

Accounts

Instructors need an account. Sign-in is by emailed link or with Google — there is no password to choose or lose. Keep access to your mailbox to yourself: a sign-in link is a credential, and anyone holding one can act as you until it is used or expires.

Students reach the service through Discord or an emailed invitation and do not choose their own account; an instructor's roster import creates it.

You must be old enough to enter a contract where you live, and at least 16.

What you may use it for

Use the service for teaching, learning and the ordinary work around them. Do not use it to:

  • break the law, or infringe anyone's copyright, privacy or confidentiality;
  • upload course material you have no right to redistribute;
  • harass, defame or target anyone;
  • attack, probe, overload or circumvent the limits of the service, scrape it, or resell access to it;
  • attempt to make the model produce content any of the above would cover.

We can remove content and suspend accounts that do these things.

Instructors: what you are taking on

Running a course on this service means you decide what students' conversations are used for, and it means:

  • You can read your students' conversations in full, and each time you do it is logged. Tell your students that before they start using it — do not leave them to discover it.
  • Their questions are sent to a third-party model provider as written, and are not de-identified first. If your institution's rules or your students' expectations do not allow that, this service is not suitable for your course.
  • You are responsible for the material you attach to a course and for having the rights to it.
  • Nothing expires on its own, but everything can be deleted. A course's conversations stay until somebody deletes them. You can delete a course, a project or the organization; a student can delete their own history in your course; an account holder can delete their account. A deletion is reversible for 30 days and permanent after that. See the Privacy policy for what a deletion does and what it leaves behind.
  • We can reach everything the service stores for your course, because we run the server and the database it lives on. We offer you no privacy guarantee against that, and neither should you offer your students one. The Privacy policy says so in full.
  • Do not use the service with students under 13, and tell your students not to type information about their health, immigration status, or sexual orientation or gender identity into it. The service stores what it is told and keeps it until it is deleted; deleting afterwards is not the same as never having sent it to the AI provider in the first place.

Student records, FERPA, and whose decision this is

We make no representation that the service is suitable for education records, or that using it complies with the Family Educational Rights and Privacy Act (FERPA, 20 U.S.C. § 1232g; 34 CFR Part 99), with your institution's policies, or with any other law governing student records. Whether it is suitable is your institution's determination, and it is not one we can make for you.

Unless we have signed a written agreement with your institution saying so, we are not acting as a "school official" with a "legitimate educational interest" under 34 CFR § 99.31(a)(1)(i)(B): we are not under your institution's direct control with respect to education records, and no institution should treat us as though we were. That designation is something an institution makes in its own annual FERPA notification. It is not something we can give ourselves.

If you bring students to this service, you confirm that you may do so on behalf of your institution and, where it requires review or approval of third-party software, that you have obtained it; that you have decided the service is appropriate for the data you will put into it; that you have given your students the notices your institution and applicable law require and obtained any consents they require; and that you will not upload anything you have no right to disclose to us. That determination is yours to make.

Nothing in these terms is a consent given on behalf of any student. Consent under FERPA can only be given by the student, in writing, to their institution, naming the records, the purpose and the recipient (34 CFR § 99.30) — and we do not collect it. What you accept by using the service you accept on your own behalf: that the Privacy policy describes indefinite retention, no per-student deletion, students' questions sent un-redacted to a third-party model provider, and instructor and operator access to what is stored.

An institution that needs a written data protection agreement, a FERPA addendum, or a completed HECVAT before adopting the service should write to privacy@wonkledge.com before its students start using it. An institution that needs something deleted — one student's records, a course, or everything belonging to it — may ask at the same address, and the Privacy policy describes what the service can delete on its own.

AI-generated content

Answers are produced by a statistical model. They are frequently wrong. The model can invent facts, misstate a deadline, misread your course material, and answer confidently either way. Nothing it produces should be relied on without checking, and it must not be treated as authoritative about a syllabus, a grade or a policy. We make no warranty as to accuracy, originality, or fitness for any purpose.

Your content

You keep ownership of what you create and upload. To run the service we need permission to do the obvious things with it: store it, process it, transmit it to the model and infrastructure providers the service is built on, and render it back to you and to the students in the relevant course.

Third-party services

The service depends on others — a model provider, Discord, a mail relay. Their terms apply to their parts, we do not control them, and an interruption at any of them interrupts the feature that uses it.

Cost controls

Model usage costs money, and the service meters it: each course has a spending cap and a daily request limit. When a cap is reached, the service stops answering for that course until the period resets or the cap is raised. That is the intended behaviour, not a fault.

Recorded costs are computed from published provider rates and are an estimate for your own budgeting. They are not a bill, and they are not guaranteed to match what the provider actually charges the operator.

Availability

The service is offered as it is and as it is available. We do not promise any level of availability. It runs on a single server; there will be downtime, defects, and mistakes that affect data. Do not make it the only route by which students can reach you, and do not rely on it during an assessment.

Ending the agreement

You may stop using the service at any time. We may suspend or end access that breaks these terms, that we are legally required to act on, or — with reasonable notice — if we discontinue the service. What happens to stored data afterwards is described in the Privacy policy, including its current limits.

Liability

To the fullest extent the law allows, we are not liable for indirect, incidental or consequential losses, for lost profits, or for lost or inaccurate content — including anything the model generated. Where liability cannot be excluded, it is limited to the amount you paid us in the twelve months before the claim. Nothing here limits liability that cannot lawfully be limited.

You agree to cover us against claims arising from material you uploaded or published in breach of these terms.

Changes

We may update these terms as the service changes. The date above says when they last changed. Continuing to use the service after a change means you accept it.

Law

These terms are governed by the laws of New York, United States, and its courts have jurisdiction over disputes arising from them.

Contact

Bloombot privacy@wonkledge.com