How can we help you?

After sign-in, the Dashboard is your home. It is customisable: add, remove and rearrange widgets, and switch presets for different workflows.

• Quick actions — jump into marking, integrity checks, students and classes.
• Jobs in progress — live status for batch marking, automarking, exam, video and essay jobs; stuck or failed jobs surface here too.
• Credit balance — personal credits, or your shared team / enterprise pool when applicable.
• Teaching widgets — at-risk students, integrity digest, assignments ready to mark, review / governance queues, question hotspots and credit runway.

New accounts also see an onboarding checklist: add students, create a class, then run one AI job. You can skip or restart it from the dashboard.

1. Open My Classes from the sidebar and create a class with a clear name (for example “Year 10 History”).
2. Optionally assign a year group so Reporting Hub and filters stay organised.
3. Add students one by one, or use bulk import to paste a list of names.
4. On each student profile you can add subject tags / labels, notes, meetings and (where used) a ULN.

Tip: keep names unique within your account. Batch ZIP folders should match roster names as closely as possible; partial matches can be linked on the results page after marking.

The first-run wizard on the dashboard tracks three milestones:

1. Add at least one student.
2. Create a class.
3. Complete one AI job (marking, integrity check, tutor session, and so on).

You can skip the wizard and return later. For a guided walkthrough of stylometric analysis specifically, signed-in users also have Getting Started under Help & Training.

Public visitors: create a free account (welcome credits, no card required) or use Try live demo to explore sample data without writing to a real account.

Pick the tool based on work type and whether you are marking a whole class or one learner:

Integrity tools (not grading): stylometric comparison, AI detector and reference checker. See the Academic integrity section below.

Batch Marking marks a whole class against a saved rubric.

1. Build or generate a rubric under My Rubrics.
2. Prepare a ZIP: one folder per student (folder name ≈ student name), with their work inside.
3. Open Batch Marking, choose the rubric, upload the ZIP and start the job.
4. When complete, review per-student results, edit feedback if needed, export or email from the student profile.

If your organisation enables governance, results may require human sign-off before release, export or email. Sign-off can be per result, per criterion or in bulk from the results page.

Use Automated Marking when you have a fixed question paper and mark scheme.

1. Go to Automated Marking → Manage Templates.
2. Upload a blank PDF/DOCX of the paper. AI extracts questions; you review marks and model answers.
3. Save the template, then run a job with student scripts in a ZIP (one folder per student).
4. Results support review, edits, student linking for partial name matches, and export.

Guided mode and review flows help when OCR or handwriting needs a second look. Reuse the same template across mocks and cohorts.

The Exam Paper Marker is for one learner at a time — ideal for short-answer tests, mocks and scanned scripts.

• Pick or create a student, upload their work (document or images).
• Optionally upload a mark scheme, or let the AI infer marking from the paper.
• The job runs through the automarking pipeline in exam-direct mode; results reuse the familiar automarking results UI.

Typical cost is about 5 credits per student (confirm in the UI). For a whole class with a reusable scheme, prefer Automated Marking with a saved template.

Use the Smart Essay Marker when you want a structured summative report for one essay.

• Paste or upload the essay and provide line-by-line criteria, or pick a saved rubric.
• The job returns criterion-level assessments (for example met / not met), overall feedback and recommendations.
• Export a PDF from the results page; recent jobs appear on the tool landing page.

Typical cost is about 10 credits per essay. For a whole class of essays, use Batch Rubric Marking instead.

The Video Marker assesses presentations, oral exams, drama performances and similar spoken tasks.

• Upload video (for example MP4, MOV) or audio (MP3, WAV).
• Provide marking criteria or a rubric.
• The system transcribes speech, then analyses the transcript against your criteria.

Large files can take several minutes. Typical cost is about 20 credits per video — confirm on the upload screen before you start.

Interactive Feedback is formative, human-in-the-loop annotation for drafts.

1. Start at Interactive Feedback, attach work to a student, and choose a rubric, pasted criteria, or general feedback.
2. AI scans the text (in ~5,000-word windows for long documents) and proposes inline comments linked to spans.
3. Add your own comments, re-run AI non-destructively, then export offline HTML when ready.

Optional reference checking and language variant settings are available when you create the session. Session creation consumes credits (see the costs reference under Account).

The AI Tutor gives formative feedback on drafts — coaching, not a final grade.

Feedback points typically include a topic, comment, concrete suggestion and a quote from the learner’s text.

Open My Rubrics to create criteria manually: assessment objectives, mark bands and descriptors that plug into batch marking, interactive feedback, the tutor and other tools.

Save rubrics to reuse across classes and cohorts. On a team, shared rubrics keep marking consistent across colleagues.

You do not need to type every descriptor by hand.

• Go to My Rubrics → Generate with AI.
• Upload an assignment brief or specification (PDF/DOCX), or paste the brief text.
• Review the extracted objectives, bands and descriptors, edit as needed, then save.

Always check generated criteria against your official mark scheme before using them for high-stakes work.

Enterprise and some team policies can require human review before AI marking results are released.

• When governance is on, results store review state (pending, signed off, invalidated after edits).
• Teachers sign off per result, per criterion or in batch from the results page.
• Sign-off stamps reviewer name and time; manual grade edits can invalidate review until re-signed.
• Export, email and profile release may be blocked until review is complete.

Organisation admins configure policy in the enterprise console; members may need to acknowledge policy before marking or reviewing.

Manage learners from Students and Classes.

• Bulk import — paste a list of names to create many profiles at once.
• Merge duplicates — combine profiles when the same learner appears twice (for example “J Smith” and “Joshua Smith”). Merges are permanent; review carefully.
• Year groups and labels — organise classes and tag students for Reporting Hub filters.

Student profiles hold marking history, authentic writing samples, worksheets, learner packs, calendar notes and meetings.

Inside a class you can create assignments, track submissions and start marking when work is ready.

• Choose how work is collected (for example teacher upload vs student upload links).
• Configure the marking mode (rubric or automated) for that assignment where applicable.
• Run AI assignment analysis for cohort-level summaries after marking (uses a small credit cost).

When an assignment uses collection mode = link, each enrolled student gets a unique URL (/submit/<token>). Students do not need a Feedback Flows account.

• Students upload a single Word document via their private link.
• Teachers monitor submissions, regenerate links if needed, download files and begin marking when ready.
• Upload typically closes once a marking job exists for that assignment.

Share links carefully — anyone with the token can upload for that student slot until the link is regenerated or closed.

On batch and automated marking landing pages, filter job history by date range and student name or ID.

On a student profile → Marking History:

• Select completed results (subject to governance release rules).
• Export XLSX — simple or advanced sheets with chosen columns.
• Email results — send to the address on the student’s Profile tab (max 50 results per send).

Unreleased governance-blocked results are skipped from export and email until signed off.

Open Students → Reporting (Reporting Hub) for cohort-level insight:

• Risk levels, integrity signals, marking averages, year-group and class breakdowns.
• Intervention queues and AI-generated cohort summaries (small credit cost).
• CSV export and a preview of the weekly digest email.

Use filters to focus pastoral and departmental follow-up without opening every student profile.

The dashboard At-risk students widget flags learners based on configurable triggers, for example:

• Any failing result,
• Average below a threshold,
• A count of fails within a window.

Triggers use rubric and automarking results linked to classes and assignments. Adjust preferences from the students list / at-risk filter controls so the widget matches your school’s intervention rules.

Upload a curriculum document or paste a topic list into the Syllabus Creator. Set weeks and lessons per week.

You get a structured plan with week/lesson numbers, topic titles, learning objectives and key outcomes. Use Generate Content next to a lesson to produce lesson plans and homework for that day, including differentiation where available.

From a student profile, open the Worksheet Generator. Choose weaknesses (often suggested from past marking), format (short answer, MCQ, and so on), difficulty and question count.

A background job builds the worksheet; download the PDF (with answer key) from Worksheets. Typical cost is about 3 credits per worksheet. Generation is disabled in demo mode.

Learner Packs turn completed marking into personalised assessment PDFs.

1. Build a template in the visual block builder (headings, rich text, logos, student fields, results tables, AI write-ups, data tables).
2. Create from a starter, build manually, or import layout from a PDF where supported.
3. Generate against one or more completed rubric or automarking results; a background job merges data into a downloadable pack.

Team-shared templates appear when you are on a team. Typical generation cost is about 3 credits per pack.

Operational emails are controlled from profile notification preferences (separate from marketing consent):

• Job complete — when rubric batch, automarking, exam, video, smart essay, AI detector or worksheet jobs finish.
• Low credits — daily check when personal balance is low (skipped for enterprise/team pool users; rate-limited).
• Weekly digest — Monday summary built from Reporting Hub-style cohort data (rate-limited).

Each email includes a one-click unsubscribe for that preference.

Stylometry compares linguistic style — useful for authorship concerns and contract cheating investigations alongside other evidence.

1. Build a portfolio — upload 3–5 authentic samples on the student profile (in-class or known handwriting).
2. Run a comparison — upload the new piece.
3. Interpret the score — a low similarity score suggests the style differs sharply from proven work; treat as a signal, not a verdict.

The AI Detector estimates the probability that text was machine-generated. Use it as one input in an investigation, together with stylometry, classroom knowledge and professional judgement.

Typical cost is about 1 credit per check. Results and history appear in the tool and can surface in integrity digests on the dashboard.

The Reference Checker runs two levels of analysis:

• Basic — match in-text citations to the bibliography; flag missing references.
• Semantic (advanced) — judge whether the cited source plausibly supports the surrounding claim (helps catch hallucinated references).

You can also enable reference checking when starting an Interactive Feedback session.

Run stylometric checks across many students at once via Batch Upload: package work in a ZIP with one folder per student, start the job, then review similarity signals from the results list.

Use the same naming discipline as batch marking so folders match roster names.

One credit is the unit for AI actions. Defaults (always confirm in the UI before you run a job):

• Batch rubric marking ~10 / student · Exam Paper Marker ~5 / student · Smart Essay Marker ~10 / essay · Video ~20 / file
• Interactive Feedback ~3 / session · AI Tutor ~2 / session · Learner Pack ~3 · Worksheet ~3
• Stylometric comparison ~1 · AI Detector ~1 · Reference Checker ~5 · Reporting summary ~1

Buy packs on Buy Credits, or use Teacher / Pro subscriptions where offered. See Pricing for current packages.

Purchase credit packs securely via Stripe on the Buy Credits page. Payment history and subscription management live under Profile (Stripe Customer Portal for plan changes and cancellation).

Referrals: share your unique link from the Referral page. When a colleague signs up and buys credits, you both receive a bonus (terms shown on that page).

Create a Team to centralise billing. The owner manages invites, access and monthly per-member limits; members draw from the shared pool and can top up when allowed.

The Team Hub shows usage analytics, member activity and (where enabled) review / compliance helpers for lighter governance without full enterprise.

Enterprise orgs get an organisation console with roles such as Owner, Manager, Quality, Commercial/Finance and Member.

• Branding, department structure, shared rubric libraries and automarking policy.
• Audit trails, compliance exports, ULN support and governance queues.
• Analytics assistants and leadership reports where licensed.

Teachers continue day-to-day marking in the normal app; admins govern from the enterprise dashboard.

The public API Hub documents REST endpoints, OpenAPI download and a Postman collection, plus live endpoint testers.

To integrate programmatically (for example with an LMS), contact us or open a support ticket to request an API key from a Feedback Flows administrator. Treat keys as secrets; rotate if exposed.

On Profile you can enable options to anonymise student names in AI prompts (placeholders in model calls; real names stay in your UI).

GDPR-oriented tools include downloading your data and requesting account deletion where available. Legal documents (privacy, DPA, sub-processors, AI usage guide) live under the Legal hub in the site footer.

Most batch errors come from ZIP structure.

Use one folder per student, named to match the roster:

MyClassSubmissions.zip
├── John Smith/
│   └── Essay.docx
├── Jane Doe/
│   └── Essay.pdf
└── ...

Full names work best. Partial names can often be linked on the results page after the job runs. Avoid nested ZIPs or a single flat folder of files with no student folders.

Check the dashboard widgets for jobs in progress, stuck and failed items. Open the job results page for the error message.

• Very large PDFs, password-locked files or corrupt uploads often fail extraction.
• Insufficient credits stop jobs before processing.
• Retry after fixing the file, or contact support with the job ID from the URL.

Supported formats vary by tool (typically PDF, DOCX, images for exam scripts, video/audio for the video marker). If a PDF is password-protected, unlock it before upload or use the unlock helper where the UI offers it.

Scans should be readable (avoid heavy skew or ultra-low resolution). Split huge multi-student PDFs into per-student files inside the ZIP.

Use Forgot password on the login page to reset via email. Check spam if the message does not arrive. Confirm you are signing in with the same email used at registration.

Enterprise invites must be accepted with the invited address; a mismatch shows a dedicated error page. Contact support if you still cannot access after a reset.

The live demo loads sample students, jobs and feedback so you can explore the UI. Many write actions (creating tickets, generating new worksheets, real marking spends) are blocked or view-only.

Create a free account when you want to mark your own work with welcome credits. Book a demo via Contact if you want a guided walkthrough for your school or ITP.