Private beta · invite only

Spot the holes in a witness statement in 90 seconds.

Conectir runs SCAN analysis on the original-language statement, flags the markers with their innocent alternatives, and produces a court-ready report. Investigative leads, never a verdict.

Request early access

Private beta · invite only · we're onboarding testers now

conectir.com/cases/c-2026-0142
ConectirCases · Search · Reports
Cases · C-2026-0142

Warehouse theft, 15 May 2026

Lead · A. van StadenStatus · In progress2 contradictions flagged
STATEMENTS
  • SN
    Susan Nyathi
    Warehouse Manager
    contradicts statement #2
  • PM
    Petros Mthembu
    Night Guard
    contradicts statement #2
  • LK
    Lerato Khumalo
    Shift Supervisor
  • JO
    James O'Brien
    Owner

BUILT FOR THE PEOPLE WHO RUN INVESTIGATIONS

  • Police detectives
  • Insurance fraud teams
  • Private investigators
  • Corporate compliance & HR
  • Defence lawyers
  • Government regulators

COMPLIANCE & ADMISSIBILITY

Built so your evidence holds up in court.

Original-language analysis, an innocent alternative for every marker, and a court-ready audit trail on every case.

  • 🇪🇺
    GDPR & UK GDPR
    Aligned
  • 🇿🇦
    POPIA
    Compliant
  • SOC 2
    SOC 2 Type II
    Audit underway
  • 27001
    ISO/IEC 27001
    On roadmap
  • CJIS
    CJIS-aligned
    Architecture review
  • ⚖
    Court-ready audit trail
    Shipped

Documentation, DPAs and security questionnaires are available to enterprise customers under NDA. Request the trust pack →

THE WEDGE FEATURE

Cross-statement reconciliation, in any language.

Drop in two to ten witness statements in their original language. Conectir builds one merged chronological timeline, then flags every place two statements disagree with each other — with the verbatim quotes from both sides.

Every flagged contradiction comes with at least one innocent alternative — translation drift, memory error, cultural framing — so the investigator decides what to follow up on. Conectir never decides who is lying.

See a worked example →
TIMELINE · WAREHOUSE THEFT
  1. 21:40Last delivery van departs (S. Nyathi).
  2. 22:15Roller door locked (P. Mthembu).
  3. 22:50Statement #1: alarm armed. Statement #2: alarm NOT armed.
  4. 01:10Motion sensor triggered, east bay.
  5. 06:30Theft discovered on opening (L. Khumalo).

SCAN STATEMENT ANALYSIS

Two ways to read a statement. You pick.

CALM

Mina ngangihamba phakathi kwesitolokanye nomama wami. Sasifuna ukuthengaisinkwa. Kwafika indoda ehlokomayo yathi siphume. Mina ngabaleka ngashiya umama wami lapho.

Bottom line: five markers flagged. Original language preserved. English gloss available on hover.

RAINBOW

Mina ngangihamba phakathi kwesitolo kanye nomama wami. Sasifuna ukuthenga isinkwa. Kwafika indoda ehlokomayo yathi siphume. Mina ngabaleka ngashiya umama wami lapho.

Same five markers, colour-coded by category: pronoun · repetition · missing info · hedge · time gap. Useful when training.

Calm is the default for production case work. One signal colour means every marker reads with the same urgency: needs review. Your eye stops asking “what colour was that one again?” and starts asking the real question — why is it flagged?

Rainbow is for teaching, training and self-review. The per-category palette makes the SCAN method visible while you build the eye for it. Toggle between them per statement.

PRIVACY BY ARCHITECTURE

Designed for material a defence lawyer will read.

  • Pseudonymisation before AI

    Names, places, ID numbers and dates are replaced with stable tokens before any prompt leaves the workspace. The AI never sees who or where.

  • Zero retention with AI

    We use Anthropic's zero-retention API mode. Your case material is never used to train models and is purged from the API edge after the request completes.

  • AES-256 at rest

    All case data is encrypted at rest in the customer's chosen region. Keys are rotated quarterly. Database snapshots are encrypted too.

  • Audit log per case

    Every read, write, AI call and report export is logged with a hash-chained timestamp. Exports include the verification URL so a defence lawyer can confirm provenance.

PRICING

Calm pricing. No surprise bills.

Prices in USD. We auto-display ZAR, GBP, EUR, AUD and CAD when we detect your region.

  • PRACTITIONER
    $76/ month
    Billed annually · $912 / year
    Save $156 / year vs monthly

    Solo investigator. Unlimited cases, 100 AI credits / month.

    • Unlimited cases
    • 100 weighted AI credits / month
    • SCAN, timeline, cross-case reconciliation
    • Court-ready reports with audit log
    • Light + dark · mobile-friendly
    Request access
  • Most chosen
    TEAM
    $297/ month
    Billed annually · $3564 / year
    Save $624 / year vs monthly

    Five seats including the team lead. 1,000 credit pool, allocated by the lead.

    • 5 seats including team lead
    • 1,000 shared AI credit pool / month
    • Team dashboard + case assignment
    • Reviewer role with approval gate on reports
    • White-label print (your logo · your signature)
    Request access
  • ENTERPRISE
    Talk to us

    Annual contract. Custom seats, custom credit pool, custom SLAs.

    • Custom seats and credit pool
    • SSO · SCIM · data residency
    • Per-tenant audit log export
    • Named onboarding lead
    • 99.5% uptime SLA
    Talk to us

Heavy month? Optional AI credit top-up packs are available inside your workspace — purchased on-demand, no auto-renewal, no surprise bills.

Most investigators find their first contradiction within 30 minutes.

Conectir is in private beta. Request early access and we'll be in touch.

Request early access

Private beta · invite only · we're onboarding testers now