Every instrument. Every score. Every quote. Every observation. Timestamped.
The clinician picks the tests, scores the tests, conducts the interview, and renders every diagnostic decision. Psygil keeps the record of all of it.
Decision gates keep you in charge of all diagnostics
The diagnostic assistant maps your evidence to DSM-5-TR criteria and surfaces convergent and contradictory findings across test data, interview, and collateral. It does not diagnose.
You render, defer, or reject each candidate, with signed reasoning in the audit log. The writing assistant will not draft a report until you have rendered at least one diagnosis.
Six stages. Gates between every one of them.
Onboarding, Testing, Interview, Diagnostics, Review, Complete. A case cannot reach Diagnostics until your testing is documented. A report cannot publish until you have signed every diagnosis.
The ingestion assistant reads referrals, prior evaluations, medical records, and court documents in PDF, DOCX, RTF, CSV, and Markdown. Bulk ingest handles fifty files or more with per-file retry so records keeping stays out of your way.
You pick the tests. You score the tests. Psygil keeps the record.
Clinical battery (MMPI-3, PAI, WAIS-IV, BDI-II, MoCA) and forensic instruments (TOMM, MSVT, PCL-R, ECST-R, MacCAT-CA, HCR-20, Static-99R). Psygil does not score from answer sheets.
Psygil v1.0 ships with twelve forensic and clinical score-entry workflows. v1.1 expands to forty-three instruments with normative context across forensic, neuropsychological, trauma, and ADHD/Autism batteries. Full inventory in the Feature Catalog.
Validity scales flag automatically. F-r, FBS-r, symptom-validity cutoffs, PAI inconsistency indices. Findings feed the evidence map next to your interview and collateral.
Charts that read the way clinicians read
Each instrument shows a normative visualization tuned to how the instrument is interpreted. MMPI-3 renders clinical-scale curves. PAI shows elevation profiles. WAIS-V shows percentile bands across index scores. TOMM and MSVT render cutoff thresholds. Validity flags color the chart when thresholds fire.
Live transcript. Structured observations. Nothing leaves the workstation.
Local Whisper transcription streams in 250-millisecond segments. No audio leaves the workstation. You pause, resume, and annotate inline without interrupting the recording.
A parallel observation rail captures appearance, psychomotor, eye contact, affect, speech, and rapport. Each field maps to a standard Mental Status Examination section, time-stamped against your transcript.
Every quote in the final report is tied to a time-stamped line. Cross-examination reads from the record, not around it.
Nine categories. Every flag tied to a paragraph.
The review assistant flags speculative language, unsupported conclusions, Daubert and Frye vulnerability, factual inconsistency, overstatement, missing caveats, source gaps, and diagnostic overreach. Severity-coded, paragraph-pinned.
You decide what to revise. The assistant does not rewrite unless you ask.
Your voice. Your vocabulary. Your signed diagnoses.
The writing assistant drafts the report in your voice and style, using your vocabulary, from the diagnoses you have approved and signed. Every section is labeled generated or draft with a confidence score. Sections awaiting upstream stages show placeholders, not prose.
The peer consult you always meant to do, structured.
Export a redacted review package. A trusted colleague imports it into their own Psygil instance, leaves paragraph-anchored comments with severity grades, records consensus or dissent, and exports a signed response.
You import their response, resolve each comment to consensus or override with attestation, and the entire exchange is written to the audit trail. No inboxes. No shared drives. No raw case record in transit.
The scoring specialist that never skips a validity check.
Between Testing and Diagnostics, a dedicated psychometrician assistant reviews every score, surfaces validity concerns, interprets against normative context, and produces an evidence-grade brief that your diagnostic workflow can cite.
Diagnostics cannot advance until the score review is complete. You still pick the tests, enter the scores, and decide what the pattern means.
Names never leave the workstation.
Local Presidio and spaCy redact all 18 HIPAA Safe Harbor categories into single-use Universal Non-Identifying Descriptors (UNIDs) before any text leaves the workstation. UNIDs are rehydrated locally on return.
Case data lives in a SQLCipher database keyed with Argon2id, stored in the OS keychain (macOS Keychain, Windows DPAPI, Linux libsecret). Nothing caches to cloud, browser, or temp.
You sign. The report locks. The package is ready.
Attestation captures your typed signature, statement, and date. Psygil hashes the final report with SHA-256, records the hash in the audit trail, archives drafts, and locks the DOCX and PDF to read-only.
Reopening a sealed case requires a supervisor override that is itself audited. The exported package meets licensing, payer, professional, and HIPAA documentation standards, and stands up in court.
Every decision, reproducible on demand.
Every action writes an immutable entry to an append-only audit log, protected by SQL triggers and chained with SHA-256. Any post-hoc tampering is detectable via the built-in chain verification tool.
On request, you export one package: intake, scores, interview, documents, your decisions with rationale, revisions, attestation, and file hashes. Review reads the record rather than reconstructing it.
The testimony package exports as a single zip: final report, integrity hash, chronological audit log, attestation record, raw test protocols, interview transcript, and every source document cited. One file, reconstructable on discovery, readable by opposing counsel's expert.
Assessment library expansion.
Psygil v1.0 focuses on the score-entry and evidence-record surface. v1.1 expands the instrument catalog and adds chart metadata so the testing record reads with the same clinical structure you use in practice.
| Group | v1.0 | v1.1 target |
|---|---|---|
| Forensic | TOMM, MSVT, PCL-R | ECST-R, MacCAT-CA, HCR-20v3, Static-99R |
| Clinical | MMPI-3, PAI, BDI-II, MoCA | Expanded mood, trauma, and personality context |
| Neuropsych | WAIS-IV score entry | WAIS-V, WMS-IV, D-KEFS, CVLT-3, Trails |
| ADHD/Autism | Template foundation | Conners-4, CAARS, ADOS-2, SCQ, RAADS-14 |
Writer drafts. Editor checks. The clinician signs.
Every draft revision, every change, every signature is attested and written to the audit log.
Watch a full evaluation run in three minutes
The demo walks through every stage from referral to attested report. The gate is visible. The decisions are visible. The chain of evidence is visible.