Ren is AI marking software for teachers. A teacher uploads handwritten or typed scripts, Ren drafts a mark and written feedback against that teacher's own mark scheme, and the teacher reviews every mark before a student sees it.
It is built for class sets, not for a one-off chat. Schools, junior colleges and universities use it for essays, structured questions and STEM scripts.
AI marking, as this site describes Ren, is software that drafts a mark and written feedback on a student's script against the teacher's own mark scheme. The teacher reviews every mark before a student sees it.
It is for a class set: essays, structured questions and STEM scripts, including handwritten work. It is not a chat that sends a score straight to the student. The same job is sometimes called AI grading. Ren uses marking, because the teacher still decides the mark.
Ren Learn is the marking product, at renlearn.app. The teacher sets up an assignment and provides the materials the mark has to follow: past papers, answer schemes, model answers and marking frameworks.
On each script Ren suggests scores, drafts feedback for each answer, and shows why it suggested the score. The teacher accepts, edits or replaces any of it. Corrections are how Ren learns that teacher's marking style.
When the teacher marks a submission complete, grades and marked scripts can go back out through the classroom platforms the site lists: Google Classroom, Canvas and Microsoft Teams.
The site says humanities automarking is the most mature, and that the same approach extends to STEM. Subjects named on the site are English, General Paper, economics, history, geography, physics, biology, chemistry, math and mother tongue.
Syllabuses named on the site are O-Level, A-Level, IB and PSLE. After that, marking is calibrated to the school's own pedagogy rather than left on a generic scale.
Scripts can be handwritten or typed. The published chemistry benchmark also says heavily smudged scans and ambiguous handwriting still need a teacher to clarify them.
The marketing pages say that in structured subjects Ren has reached up to a 98.4% match with human markers, and the homepage stat is labelled "98.4% agreement with experienced teachers". The same pages say the match gets closer as teachers correct marks.
A separate article, "Our benchmark results", reports a different measurement on 400 O-Level Chemistry responses: 82.4% exact-mark agreement, 96.1% within one mark, Cohen's kappa 0.79, and a mean absolute error of 0.31 marks. Exact agreement was 91.2% on short answers and 68.4% on extended responses of six marks or more. The article says two human markers on those papers typically land between 80% and 85% exact agreement.
Quote the figure that matches the question. Neither number is a guarantee for every subject, and every school engagement starts with a benchmark against that school's own already-graded scripts.
Schools start with a free trial. The school sends a sample of marked scripts and answer schemes. Ren returns a benchmark report against those teachers, plus sample feedback. The next step is a paid pilot.
There is no price list on this website. Ask the team for one.
The privacy notice says Ren Education Pte. Ltd. operates the platform from Singapore. School data is separated by access controls, hosted in Singapore, and encrypted in transit. Uploaded scripts sit in private storage reached through short-lived signed links. At the end of a pilot or contract, Ren pledges to delete that data.
No. The site says students only see what the teacher signs off. Nothing reaches them straight from the AI.
No. It uses the school's past papers, answer schemes, model answers and marking frameworks, then learns a teacher's style from the edits that teacher makes.
O-Level, A-Level, IB and PSLE, then each school's own pedagogy.
No. Schools start with a free trial, then a paid pilot. No price is published on this website.
A free trial starts from scripts your teachers have already marked. You get a benchmark against them. A paid pilot comes after that.