AI forensic data analysis & research laboratory

Messy evidence. Reviewable conclusions.

YY examines datasets, operational records, and AI-system behavior. We return source-linked findings, explicit gaps, and inspectable artifacts—then state where the evidence stops.

50% off AgentSafe and Dataset Preflight through .

Lead product

Dataset Preflight

Regular price $39Sale price $19.50one-time

Inspect a dataset locally, catch duplicate, provenance, and secret-risk findings, then optionally export selected findings and policy checks for CI.

  • Runs locally
  • No network requests
  • Checkout via Polar

AgentSafe

Regular price $49Sale price $24.50one-time

Prove approval gated your OpenAI Agents SDK FunctionTool before it ran, with deterministic local evidence.

Paid services

Bounded operational cleanup, priced upfront

Two fixed-scope pilots for non-developer teams. Review the exact input, output, timing, price, and boundary before sharing files or paying. Browse all services.

First paid pilot / $149 / fit check first

Calibration Register Cleanup

A bounded cleanup service that turns one existing gage list and up to 25 supported calibration-certificate PDFs into a source-linked register and an explicit review queue.

  • One site
  • Up to 25 supported PDFs
  • Three business days

Data-entry assistance; not calibration, metrology judgment, or compliance advice.

First paid pilot / $19 once

Association Minutes Draft

A draft of board minutes, motions, votes, decisions, and action items that is checked against one authorized association-meeting recording before delivery.

  • One authorized recording
  • Up to 60 minutes
  • Draft by end of second business day

A recording-checked draft; not adopted minutes, legal advice, or a complete record of events outside the recording.

Available now

Useful products, priced plainly

Two one-time founding licenses and one metered Poe utility. Each exposes its measured scope, price, and non-claims before payment. Compare the paid products.

Lead product · Regular price $39Sale price $19.50once

Dataset Preflight

Inspect a local dataset for duplicates, provenance gaps, and secret-risk signals, enforce your thresholds, and emit opt-in SARIF and JUnit evidence for CI.

  • v0.2.0
  • SARIF 2.1.0 + JUnit
  • No network requests

A bounded evidence packet—not a safety, license, or training-readiness verdict.

Founding release candidate · Regular price $49Sale price $24.50once

AgentSafe

A deterministic, local evidence checker for one specific behavior: approval-to-execution in OpenAI Agents SDK 0.19.x FunctionTools.

  • OpenAI Agents SDK 0.19.x
  • Runs locally
  • Deterministic JSON evidence

One measured FunctionTool behavior—not a security audit or certification.

Metered Poe utility · 500 YY points + Poe base

Chat Receipt

Export the conversation visible to the bot as a deterministic ZIP with a readable transcript, canonical JSON receipt, SHA-256 manifest, and local verifier.

  • 501 points per message at publication
  • No downstream model call
  • Deterministic ZIP + verifier

Exports only the history Poe presents to the bot. Verification proves internal consistency—not truth, identity, origin, or completeness.

Free and open work

Start with something inspectable

Use a browser utility, review a specification through two named models, or install a released package.

  • Open research · Python + RTL

    Analog Matmul

    A behavioral mixed-signal matrix model, held-out 4 × 4 characterization, driver contract, and open-silicon roadmap for inference and hybrid-training research.
  • Free · Poe

    Spec Court

    Send one bounded software or product specification to two named reviewer models and download their independent findings as a deterministic, locally verifiable receipt.
  • Free · browser

    Evidence Verifier

    Recompute every SHA-256 digest in a supported evidence packet and inspect the result without uploading the files.
  • Open source · Node.js

    ClipCheck

    Offline, deterministic preflight checks for short-form video files before they are uploaded.
  • Open source · Python

    Workload Receipt

    Deterministic, content-free receipts for observed AI workload usage metadata, with explicit evidence gaps and later integrity verification.
  • Open source · Python

    Corpus Evidence

    An open-source Python library that inventories a local corpus, incorporates optional provenance declarations, and produces a deterministic manifest that can be checked for drift later. Alpha supports CPython 3.11 on Windows only.
  • Open source · Zig

    Chi

    Streaming, checkpointable Zig components for bounded corpus ingestion, text cleanup, document filtering, and exact or approximate deduplication.

How we build

The same contract across every surface

Services, software instruments, open tools, and research prototypes are different outputs of one laboratory contract: preserve the evidence, expose uncertainty, and keep every conclusion inside its supported boundary.

  1. Determinism before convenience

    For deterministic tools, the same supported inputs should produce the same bytes. Results should be reproducible, not anecdotal.

  2. Local by default

    A file-inspection tool should be useful on one machine, offline, before it asks anyone to send data somewhere else.

  3. Only claim what we can show

    No unmeasured numbers or implied guarantees. Every product page states the measured scope and the boundary around it.

  4. Legible artifacts

    Outputs use readable formats, published schemas, manifests, or source so another person can inspect the result.

Contact

Reach YY directly

This site has no contact form, newsletter, account system, or analytics. Email reaches the company directly.

Privacy
privacy@yycore.ai
Security
security@yycore.ai
Notices
notices@yycore.ai