GhostWell AI90-second tour

AI EVIDENCE SYSTEM FOR ORPHAN-WELL REVIEW

Find what the map forgot.

GhostWell reconciles historical maps, aerial imagery, and public records to help infrastructure teams prioritize safe field verification.

Model proposes. Specialists verify. Policy decides. Humans authorize. Trace proves.

Start the 90-second tourView a worked case

Synthetic, generalized demonstration. No operational location disclosed.

Conceptual illustration transitioning from a historical map into an aerial landscape with a highlighted evidence-review ring; it is not site evidence.

Conceptual illustration — not site evidence.

THE OPERATING MODEL

Evidence advances through explicit boundaries.

  1. 01

    Reconcile evidence

    Align historical maps, aerial imagery, and public records.

  2. 02

    Inspect and challenge

    Bounded specialists inspect independently; typed validation verifies evidence integrity, and the Skeptic produces the strongest alternative.

  3. 03

    Apply policy

    Deterministic rules expose inputs, thresholds, rationales, and the bounded disposition.

  4. 04

    Authorize and prove

    A human records a separate scoped authorization event; the trace preserves the sequence.

Synthetic, generalized demonstration

Ledger Gap 07

Allegheny County, Pennsylvania

Synthetic workflow scenario with a strong archive mark, a plausible modern surface cue, and no documented registry record inside 500 meters.

Recorded disposition
FIELD_VERIFICATION_RECOMMENDED
Registry relationship
NO_DOCUMENTED_MATCH_WITHIN_100M
Skeptic's strongest alternative
Small vegetation or grading artifact

Recommendation is not authorization. A human must separately authorize any field action.

View the worked case

DISCOVERY / TRUST BOUNDARIES

Questions judges ask

Concise answers grounded in the same public claims used by the product, metadata, and machine-readable summaries.

01

What is GhostWell AI?

GhostWell AI is an evidence system for orphan-well review. It reconciles historical maps, aerial imagery, and public records so infrastructure teams can prioritize safe field verification.

02

What evidence does it reconcile?

The demonstration aligns synthetic historical-map and aerial evidence with generalized public-record context, then preserves source provenance, limitations, challenge, policy, and trace.

03

Does GhostWell confirm a well or detect methane?

No. RGB imagery cannot confirm a well or detect methane or contamination. GhostWell produces bounded evidence observations; its strongest machine disposition is FIELD_VERIFICATION_RECOMMENDED.

04

Can the AI authorize field work?

No. Models cannot authorize field work. Deterministic policy produces a bounded disposition; any field action requires separate human authorization plus program access and safety controls.

05

What is the difference between Recorded Demo and Live GPT-5.6?

Recorded Demo is deterministic, credential-free, and incapable of calling OpenAI. Live GPT-5.6 is server-only, opt-in, schema-validated, timeout-bounded, and fails closed without weakening Recorded Demo.

06

Are the displayed cases and locations real operational sites?

No. Every displayed case is a synthetic, generalized demonstration. County and state labels are contextual only; no operational location, field destination, or documentary site evidence is disclosed.