VeriScript

Inspiration

Academic integrity tools often focus on the submitted document after the writing is complete. Turnitin’s core workflow compares submitted text against web pages, publications, and previous submissions, then provides similarity and AI-writing indicators for educator review. Aidify takes a more process-oriented approach by tracking activity, time, revisions, and playback in Google Docs through a browser extension.

These tools inspired a broader question: what if the writing process could be preserved inside the writing environment itself?

VeriScript was created to give students and lecturers a more complete view of how academic work was produced. Rather than relying only on a final similarity score or limiting process evidence to Google Docs, VeriScript combines document creation, process capture, verification, and human review in one platform.

What It Does

VeriScript is a local-first academic writing and review platform.

Students can:

• Write assignments in a rich-text editor.
• Format documents with headings, fonts, lists, images, tables, and links.
• View drafts, assignments, and writing metrics from a student dashboard.
• See keystrokes, revisions, pauses, paste events, and Process-DNA signals.
• Export Word, PDF, or text documents with embedded verification data.

Lecturers can:

• Load or upload student submissions.
• Verify Process-DNA signatures.
• Review writing metrics, similarity signals, citation coverage, and AI-likelihood indicators.
• Replay the writing process event by event.
• Inspect paste context, document history, and source-related evidence.
• Record a human review decision and supporting notes.

Turnitin is designed primarily for large-scale similarity and submission analysis. Aidify focuses on observing work performed in Google Docs. VeriScript connects document creation and process evidence in a dedicated academic writing environment that supports rich documents, native exports, citations, images, tables, and lecturer review.

VERISCRIPT DOES NOT MAKE AN AUTOMATIC MISCONDUCT JUDGMENT. IT GIVES EDUCATORS BETTER EVIDENCE FOR A FAIR CONVERSATION.

How We Built It

We built VeriScript as a React and TypeScript application with a native Tauri desktop layer.

The main technologies include:

• React and TypeScript for the interface.
• TipTap for the rich-text writing experience.
• Zustand for application, assignment, authentication, and session state.
• Rust and Tauri for native desktop functionality.
• SQLite-backed Process-DNA session recording.
• HMAC signatures for tamper-evident document verification.
• DOCX, PDF, and text export with embedded evidence.
• OpenAI and NVIDIA-compatible analysis providers for optional AI-assisted review.

We deliberately chose a dedicated writing environment rather than building only a browser extension. This allows VeriScript to capture structured writing events, preserve document snapshots, verify exported evidence, and provide a consistent experience across the student editor and lecturer review workspace.

Codex and GPT-5.6 were used throughout the entire build. GPT-5.6 helped us research the academic-integrity problem, compare existing systems such as Turnitin and Aidify, assess gaps in the original product, prioritize improvements, make architecture decisions, implement features, debug failures, review screenshots, validate exports, and prepare the final demonstration workflow.

Challenges We Ran Into

One major challenge was reconstructing the document during playback. Early versions used sparse HTML snapshots, which caused the document to appear blank and then load suddenly. We redesigned the playback fixtures and timing behavior so typed content appears progressively and long pauses can be compressed for demonstrations.

Another challenge was balancing useful signals with responsible interpretation. Metrics such as paste events, similarity, and AI-likelihood can provide context, but they cannot prove misconduct. We designed the interface to separate observed evidence from interpretation and keep the lecturer’s judgment central.

We also had to support rich document content while preserving Process-DNA, including formatting, images, tables, exports, signatures, and native desktop behavior.

A further challenge was platform scope. Aidify demonstrates the value of process evidence in Google Docs, but VeriScript needed to support a complete writing workflow rather than depend on one external editor. This led us to build the editor, evidence capture, export, and review experience as one integrated system.

Accomplishments We’re Proud Of

• Built a working student writing environment with live Process-DNA capture.
• Added a student dashboard for drafts, metrics, assignments, and workspace navigation.
• Added replayable writing history with keystrokes, pauses, deletions, and paste events.
• Implemented signed document export and verification.
• Added similarity, citation, and AI-writing review signals with clear limitations.
• Added support for configurable OpenAI and NVIDIA analysis providers.
• Built a lecturer review workflow that combines document evidence and writing history.
• Completed native Tauri build verification.
• Produced a polished narrated product demonstration.
• Used Codex and GPT-5.6 to move from research and gap analysis to a working end-to-end product.

What We Learned

We learned that academic integrity is not just a detection problem. It is a trust, transparency, and communication problem.

A score alone is not enough. Students need to understand what is captured, lecturers need to inspect the underlying evidence, and institutions need to keep human judgment at the center.

We also learned that process evidence is most useful when it is connected to the document itself. Activity data, revision history, citations, pasted material, and final content should be reviewed together rather than treated as isolated signals.

Finally, we learned that product quality matters as much as technical capability. A strong integrity system needs a usable writing experience, reliable exports, clear review workflows, accessible explanations, and careful handling of uncertainty.

What’s Next for VeriScript

Next, we plan to:

• Add full citation management for APA, Harvard, IEEE, and other academic styles.
• Build bibliography and source-management tools.
• Improve image, table, footnote, and document-layout support.
• Add institution-managed accounts and secure cloud synchronization.
• Expand plagiarism and similarity comparison through approved licensed providers.
• Improve AI-writing analysis with transparent explanations and confidence limits.
• Add lecturer feedback and student response workflows.
• Provide richer draft history, version comparison, and collaborative review.
• Expand accessibility, mobile responsiveness, and cross-platform native builds.
• Introduce institution-level privacy, retention, consent, and audit controls.

Built With

• react
• rust
• sqlite
• tauri
• typescript
• zustland