Introducing Vibe Simulation: Interactive Simulations You Can Trust

If you've ever nudged a slider and felt a concept click, you already get our philosophy. Vibe Simulation exists to make hard things simple, and to make results you can trust.

• Interactive: you steer parameters and watch cause-and-effect unfold.
• Accessible: runs in a modern browser; minimal clicks to insight.
• Verifiable: every meaningful run can produce a signed, replayable run receipt—inputs, scene/version, and key state transitions—so others can reproduce and discuss what actually happened.

This blog is the build log for that vision.

Note: PhET is a pioneering project from the University of Colorado Boulder and a major inspiration for our user-centered design and research ethos. We'll reference their work and research openly; PhET is not affiliated with Vibe Simulation.

Why simulations, why now?

Interactive simulations help learners form mental models because you see the consequences of changing variables, not just read about them. That's borne out in years of PhET research on design and learning, and in broader studies showing simulation-based training improves skills and retention.

Meanwhile, education and training are being reshaped by AI. Assessment models are shifting—more digitization, more orals/vivas, and tighter provenance around work. The direction of travel is clear: keep what's great about exams, plus richer evidence that the learner actually understands.

Translation: this is the moment to pair interactive simulations with verifiable outcomes.

What makes Vibe different

Most sims stop at a pretty graph. Vibe starts there—and then adds proof.

• Reasoning Reels (tamper-evident recordings): Every simulation now includes "Export Reel" and "Verify Reel" buttons. Click "Export Reel" to download a cryptographically signed .saxlg file containing a complete, tamper-evident record of your interaction session—including parameter changes, pointer movements, simulation measurements, and state digests.
• Hash-chain verification: Each event is chained using SHA-256, creating an immutable audit trail. The system generates ephemeral Ed25519 signatures for authenticity proof, with automatic fallback for browsers without Ed25519 support.
• Standalone verifier: The "Verify Reel" button opens our independent verification page where you can upload and validate any .saxlg file. It recomputes the hash-chain and verifies signatures, providing cryptographic proof that your simulation results haven't been tampered with.
• CPU-optimized recording: Recording uses lightweight FNV-1a hashing during interaction (~1% CPU impact) with full SHA-256 hash-chain computation only at export time. Events are throttled intelligently (20Hz pointer logging, 1Hz state sampling) for minimal performance impact.

Our browser-native provenance engine captures every meaningful interaction: parameter adjustments, mouse/touch events, computed measurements (CFL numbers, residuals, energy drift), and significant milestones (convergence events, stability warnings). The result is a scientifically rigorous, tamper-evident ledger that transforms interactive learning into verifiable experimentation.

Technical note: The system is designed for scientific reproducibility—not certification. It provides the cryptographic foundations that support audit workflows organizations already recognize, with clear technical documentation rather than compliance theater.

Under the hood is our browser-first provenance engine—purpose-built to make every meaningful run reproducible and tamper-evident. We're designing adapters to established provenance ideas (W3C PROV) and content authenticity efforts (C2PA / Content Credentials) so your evidence travels with your work.

We're not claiming certification. Our goal is to support audit/review workflows that organizations already recognize (e.g., NIST AI RMF practices, EU AI Act timelines)—with clear language, not compliance theatre.

Who Vibe is for (today)

1. Learners & teachers
   Use curriculum-aligned interactive simulations to explore, then attach the run receipt to assignments for fairer re-grading and discussion. If you love PhET's clarity, you'll feel at home—then notice the provenance layer that makes moderation easier.
2. Engineers & analysts
   Run fast "what-ifs," share a reproducible receipt with peers, and keep your decision trail clean. It's a practical way to bring explainability to everyday modeling—without extra meetings.
3. Training, safety & ops
   Scenario drills where every decision is replayable. That helps internal reviews and aligns with the push toward documented, risk-aware AI and analytics.

Where AI fits (without the hype)

AI should be a coach, not a crutch:

• Guided prompts nudge toward meaningful parameter changes.
• Targeted feedback is grounded in the sim's actual state, not generic text.
• Assessment pairing: a run receipt + a short viva (oral check) confirms understanding faster and fairer than policing prose—and matches the way many institutions are already adapting.

Roadmap highlights

• ✅ Reasoning Reels (implemented): "Export Reel" and "Verify Reel" buttons now live in every simulation with full SHA-256 hash-chain and Ed25519 signature support.
• Evidence Drawer (next): enhanced in-sim display showing session metadata, hash verification status, and exportable summaries.
• Run Receipts API: standard JSON for receipts so teams can build their own dashboards and analysis tools.
• Provenance adapters: export mappings to W3C PROV and Content Credentials for interoperability.
• Privacy & governance notes: clear statements on what we collect (and don't) with cryptographic transparency.
• Classroom kits: worksheet templates that reference run receipts for faster marking and moderation.
• Performance analytics: aggregate learning insights from verified simulation sessions (privacy-first).

How we write & build

We take cues from the best tech blogs: ship small, explain clearly, show the edges. We'll publish:

• Design notes: why a slider is here, not there—and what user research says. (PhET's research culture is a north star.)
• Validation posts: what our telemetry says about where learners get stuck (aggregate, privacy-first).
• Standards & safety: plain-English guides to NIST AI RMF concepts that matter to sim builders, and a running explainer on the EU AI Act timeline that teams ask us about.

If you're new here

• Try the new Reasoning Reels: Visit any simulation on Vibe Simulation, interact with the parameters, then click "Export Reel" to create your first verifiable physics experiment. Use "Verify Reel" to see the cryptographic validation in action.
• Explore the technical details: Check out our complete technical documentation for the SHA-256 hash-chain, Ed25519 signatures, and CPU optimization strategies.
• Bookmark our resources: Save the interactive simulations hub and our verifiable simulations documentation—we're continuously updating both with new capabilities.
• Educators using PhET: Tell us what evidence formats would make grading and moderation easier with verifiable simulations—we're listening and building accordingly.