On request
Detailed sheets and workflow overview
One-page sheets for students, faculty and institutions, and the full workflow overview, sent after you request a demo.
Request the details →How it works
From a student's AI conversation to a campus-wide picture.
CiteKeep works in two places: a browser extension where students use AI, and the LMS, where instructors and administrators already work. One record connects them.
Watch the two-minute introduction, then follow each step below.
Video transcript and description
AI use is everywhere in education, but it's invisible.
And AI detection tools are unreliable.
There's a better way. Standardized AI disclosure, not AI detection.
Introducing CiteKeep, the AI provenance platform for higher education.
Professors set clear AI expectations in your LMS: start from a preset, adjust any category across four plain-language tiers, then embed the policy in any assignment.
The Instructor Dashboard shows disclosure completion, and Class AI Use Insights shows how students engage with AI, so professors can guide them without playing detective.
Students capture and disclose AI use from any AI platform, through the browser extension or the web app. Every record carries its provenance, Content Verified and saved to the student's own vault.
Students can share a record, copy citations in five styles, and generate a disclosure statement, each linked to the verifiable record.
Students submit their records to the assignment. Disclosure, on the record, ensuring academic integrity.
Because every policy and record uses the same CiteKeep taxonomy, institutions see AI usage and disclosure patterns across every course.
One system for students, professors, and institutions.
Schedule a demo at citekeep.com. Limited pilots are available for Spring 2027.
CiteKeep. Standardizing AI disclosure for higher education.
01 · Students
Disclose AI use. Demonstrate your learning.
Students record AI use while they work, not after. The conversation and the disclosure are saved together and kept private until the student submits or shares them.
Step 1: Install and sign in
Add the CiteKeep extension to Chrome or Edge, pin it, and sign in with a school account.
Step 2: Work as usual
Use AI as you already do. On supported platforms the conversation is captured automatically, and any other tool can be added by hand.
Step 3: Check the policy
Each assignment shows the instructor's policy for all 16 uses: permitted, permitted with disclosure, or not permitted.
Step 4: Disclose
Choose the uses that apply and how much of the work AI shaped. CiteKeep writes the disclosure statement and a citation in APA, MLA, Chicago, Harvard or IEEE.
Step 5: Show your learning, then submit
Add a reflection, summary, critique or sources when the assignment asks, then submit from inside the course. Used no AI? Submit a No AI Use Attestation instead.
Disclose AI useEssay 2: Policy briefProcess and polish
Tier BResearch supportDisclose
Tier CEditing, proofreading, and formattingDisclose
Tier DDrafting contentNot permitted
Learning evidenceReflectionSummaryCritiqueSources
Content Verified
- Private by default. Records stay in the student's Vault until they share them, and share links can be revoked.
- Content Verified. Automatic captures are marked verified; manual entries are marked student-attested.
- Flagged, never blocked. A use outside the policy is still saved, and flagged for the instructor.
02 · Faculty
Support your students through guided AI disclosure.
Instructors set expectations once and read disclosures where they already teach. CiteKeep runs inside the LMS, so policies, rosters and submissions stay attached to the course.
Step 1: Open CiteKeep in the course
CiteKeep launches from course navigation through LTI 1.3, with the roster already synced. No separate login.
Step 2: Set the policy
Start from a preset: Open, Process and polish, Process only, or No AI in the work. Then change any single use.
Step 3: Insert it into the assignment
One click places the policy in the assignment description, so students see the same rules, including any required reflection, critique or sources.
Step 4: Review each submission
A record shows the disclosed uses, the generated statement, the learning evidence and the conversation itself. Automatic captures carry a Content Verified check.
Step 5: Respond and see patterns
Comment on a shared record so feedback stays with the work, and see class-level patterns by assignment. Indicators support teaching; they are not misconduct findings.
Assignment policyEssay 2
OpenProcess and polishProcess onlyNo AI
Tier ASupported my learningPermitted
Tier BSupported my processDisclose
Tier CRefined my workDisclose
Tier DGenerated content in my workNot permitted
Require:ReflectionSources
- Three policy states, set for a whole tier or any single use.
- Canvas through LTI 1.3 Advantage, with roster sync and deep linking. Moodle 4.5 LTS and 5.x in pilot.
03 · Institutions
Stop policing AI. Start normalizing honesty.
Administrators publish a baseline policy and follow disclosure patterns across every connected course. Reporting is aggregate by design: it shows how the campus documents AI use, not what any one student did.
Step 1: Publish a baseline
Set an institution-wide baseline and reusable templates. Course and assignment policies build on it, and coverage shows where a policy is in force.
Step 2: Follow adoption each term
Connected courses, faculty, active students and records by term, with disclosure coverage alongside.
Step 3: See how students disclose
Totals across the four tiers and 16 uses show where AI support concentrates, in aggregate only.
Step 4: Manage people and access
Roles and identity sources for every account. This area never shows anyone's disclosure activity.
Disclosure practicesFall term · aggregate
Courses48
With policy41
Records1,920
Illustrative data. Share of records using each tier.
- A dedicated database for each institution.
- Small groups protected: reports never single out a small group of students.
- Not a ranking: learning-evidence indicators are not grades, misconduct signals or student rankings.
What CiteKeep does not do
No keystroke logging. No screenshots. No AI detection.
CiteKeep supports voluntary disclosure of AI assistance. It records what a student chooses to disclose, and it does not monitor or detect undisclosed AI use. Disclosure, not detection.
Works with
Where your students and faculty already are.
Learning platforms
- Canvas, LTI 1.3 Advantage
- Moodle 4.5 LTS and 5.x, in pilot
Roster sync and deep linking in Canvas.
Browsers
- Google Chrome
- Microsoft Edge
Saved automatically
- ChatGPT
- Gemini
- Claude
- Microsoft Copilot
- Perplexity, and other supported sites
Any other tool can be added by hand, marked student-attested.
Citation styles
- APA
- MLA
- Chicago
- Harvard
- IEEE
Take it with you
Share it with your faculty, IT and academic integrity teams.
Web page
The taxonomy
The 16 uses in 4 tiers that every policy, disclosure and report uses.
Explore the taxonomy →Web page
Privacy
What CiteKeep stores, who can see it, and how a student stays in control.
Read the privacy policy →Spring 2027 pilot
Shape CiteKeep with us.
Pilot institutions receive full use of CiteKeep for the Spring 2027 term, onboarding completed before classes begin, and a direct role in shaping the product.