Rheton makes academic integrity visible by design — capturing the whole research-and-writing process instead of scanning finished essays for AI content.
Obline Education is building the assessment infrastructure for the AI era. Founded in 2026, pre-seed, self-funded to date. Rheton makes academic integrity visible by design — capturing the whole research-and-writing process instead of scanning finished essays for AI content.
We are raising £100,000 to fund our first institutional pilots and complete the build.
Coursework was built on an assumption: that a finished document reflects the mind that produced it. That cracked with the internet, diploma mills and plagiarism — and tools like Turnitin emerged to stop academic misconduct. Universities now face their second major challenge: spotting and stopping AI misuse. The institutional response is failing — detection is an unwinnable arms race, and the fallback of exam halls throws away everything authentic about true academic writing.
Rheton brings the complete research-and-writing lifecycle into one collaborative workspace: reading, planning, drafting, peer critique, and consensus-gated submission — with every contribution attributed and every tutor comment held in context, right where it belongs.
When the journey is visible and attributed, the work becomes its own evidence. There's no quiet gap left for a chatbot to fill, and nothing to detect after the fact. This is the constructive third option between "detect and punish" and "revert to exam halls".
For scale: Turnitin alone licenses to 16,000+ institutions and 70M+ students. Even a tenth of that footprint would make us a large company.
IP / secret sauce: durable multi-anchor annotation recovery across PDF re-renders, and per-character authorship attribution across live group editing — brought together in one platform for assessment in the AI age. Roadmap: a marking copilot that directs tutor attention and auto-flags outlier marks for second marking.
Five-plus years in EdTech watching AI dismantle traditional assessment methodology. As Head of Programme Management, I have lived the struggles that moving away from essays causes. I am building Rheton for people like me: the whole research-and-writing process in a single platform that supports both students and tutors.
CMO of Learna, practising GP and leader in medical education, with years of experience working alongside Cardiff, the University of Buckingham and the University of South Wales. Brings contacts and expertise in high-stakes courses facing elevated AI-assessment risk.
Per-institution annual licensing at £150 per assignment — each assignment covering up to 30 students, two tutors (marker and second marker) and one admin. Built on LTI 1.3, so it drops into Canvas, Moodle, Blackboard and the rest with no rip-and-replace. Grades flow back automatically into the LMS, and rosters provision themselves.
Land with a single department that's feeling the assessment pain most, prove it, and expand across the institution. Standards-based integration keeps switching costs low for the buyer and retention high for us.
| Detects after submission | Group authorship | Full lifecycle capture | |
|---|---|---|---|
| Turnitin / Clarity | ✓ detection-first | ✗ | ✗ |
| Cadmus | partial | ✗ | partial |
| Hypothesis / Perusall | ✗ annotation only | ✗ | ✗ |
| Gradescope | ✗ marking only | ✗ | ✗ |
| Rheton | nothing to detect | ✓ per-character | ✓ read → mark |
Barriers to entry: the process-capture architecture (durable annotation anchoring, attributed co-editing) took real technical de-risking and is not a quick clone. The LTI/AGS integration layer creates institutional switching costs once adopted. And the framing of visibility over surveillance is hard for detection-first incumbents to retrofit without contradicting their own positioning.
Pre-launch. Working prototypes validate the highest-risk technical bets, and the production architecture is specified.
Pipeline: conversations underway with the online university Learna; outreach in progress to Cardiff University and the University of South Wales to trial the product, with links to Cambridge and Leeds widening the pilot net.
Identify the department under the most acute assessment-integrity pressure — typically large-cohort, high-AI-use disciplines such as English — run a free or discounted pilot for one module or cohort, prove the workflow, then expand to department-wide and institution-wide licensing.
The acquisition channel at this stage is direct founder-led outreach to department heads and heads of learning technology, backed by the "why now" pressure every institution is already under.
| Revenue | Costs | Net | |
|---|---|---|---|
| Year 1 | £0 — free pilot year, by design | £114k | (£114k) |
| Year 2 | £22k — pilots convert to paid | £175k | (£153k) |
| Year 3 | £151k — land-and-expand compounds | £247k | (£96k) |
Year 1 is a deliberate free pilot year to gather evidence and reference customers. Losses narrow from Year 2 as converted pilots expand, with a breakeven trajectory into Year 4. The £100k pre-seed funds the build-and-pilot phase; a subsequent seed round funds the scale-up.
No institutional funding to date — self-funded through the prototype and business-plan stage. We're raising £100,000 to fund the initial build and pilot of Rheton. Key milestones this unlocks: