Product

From custodian map to
Bates-stamped production.

One platform across the full EDRM. Statistically certified recall. A cryptographic audit trail you can hand to opposing counsel.

Try the demo → Book a walkthrough
What's different

Defensibility is a loop, not a label.

Every eDiscovery vendor claims "defensible AI." ORCA is the only one that closes the loop end-to-end, in your infrastructure, with statistical guarantees you can defend in a 26(f) report.

1

Classify

Active-learning models trained on your corpus, your seed labels, your privilege calls. No pretrained data leaks; the model only knows what your reviewers have labeled.

2

Certify

Statistical estimators certify that you've hit your recall and precision targets — before you stop reviewing. The math, not the vendor, says you're done. Conservative lower-confidence bounds, not point estimates.

3

Audit

Every classification, certification decision, and reviewer touch is logged with cryptographic provenance. Reproducible from raw corpus to final production.

This is the loop the 7th Circuit's Pilot Plan, Sedona Conference Commentary, and Rule 26(b)(1) proportionality framework have been asking for since 2007. ORCA is the first platform that ships it as a primitive, not a manual workflow.
Coverage

One audit trail, every EDRM stage.

Most platforms cover part of the EDRM and hand off to vendors for the rest. ORCA is end-to-end — same audit trail from custodian identification to trial-pack production.

1 · Identification
Custodian Map
Auto-discovers data sources across Microsoft 365, Google Workspace, file shares, chat platforms.
2 · Preservation
Legal Hold
Issues holds, tracks acknowledgments, monitors for drift.
3 · Collection
Connectors
Forensically defensible pulls from M365, Workspace, Slack, Teams, Zoom, mobile, web, file systems.
4 · Processing
Pipeline
OCR, dedupe, near-dedupe, language detection, metadata normalization, family threading.
5 · Review
TAR 3.0 + Active Label
Continuous active learning with privilege, responsiveness, and issue tagging. Reviewer UI built for speed.
6 · Analysis
Cert Engine
Statistical certification of recall and precision per issue, per custodian. Audit-ready stop decision.
7 · Production
Bates + Load Files
Bates numbering, redactions, privilege logs, EDRM-XML, Concordance DAT/OPT, platform-loadable archives.
8 · Presentation
Trial Pack
Curated exhibit sets, deposition-ready PDFs, chain-of-custody report.
Integration

Data goes in the way custodians work. Comes out the way courts expect.

Comprehensive connector coverage on the way in. Every major load-file format on the way out. Cryptographic provenance end to end.

What ORCA ingests

Cloud productivity

  • Microsoft 365 — Exchange, OneDrive, SharePoint, Teams
  • Google Workspace — Gmail, Drive, Chat

Collaboration & chat

  • Slack workspaces
  • Microsoft Teams (deep)
  • Zoom — recordings + chat transcripts

Mobile & device forensics

  • Cellebrite-format extractions (UFD, mobile backups)
  • Forensic disk images (E01, AFF, raw)

Web & social archival

  • Hanzo, PageFreezer, and equivalent web-capture formats
  • Social media archives

Legacy & enterprise

  • File systems — SMB, NFS, S3, GCS, Azure Blob
  • Tape backups (LTO, DLT) via collection partner
  • PST, MBOX, MSG, EML
  • ~300 document formats — Office, PDF, CAD, DICOM, audio, video

What ORCA produces

Load files & exchange formats

  • Concordance DAT/OPT load files
  • TIFF + DAT productions
  • Relativity-, DISCO-, Everlaw-, Casepoint-, Nuix-loadable archives
  • EDRM XML
  • RSMF for chat productions

Document outputs

  • Bates-stamped PDF / PDF-A (or single-page TIFF)
  • Native files with Bates-stamped filenames
  • Per-recipient redaction sets

Audit artifacts

  • Privilege logs — standard + custom schemas
  • Cryptographic chain-of-custody manifest
  • TAR Disclosure Package (Rule 26(f)-ready)
Where ORCA does not go. We don't host your data unless you pick the Solo or Org tier. We don't share data with anyone, including subprocessors, outside the matter. We don't use your data to train models — yours or anyone else's.
Tiers

Pick the deployment that matches your matter.

From a 60-second anonymous demo to a fully air-gapped self-host. Pricing on the Pricing page.

Try

Anonymous demo

For anyone
  • Public corpus, no signup, no data uploaded
  • See cert math + audit trail on real-but-public data
  • 60-second access
Solo

Hosted by PlumGen

For solo practitioners and small firms
  • Per-matter pricing, no annual commitment
  • Data hosted in PlumGen's Google Cloud or Azure tenant; region you select
  • Zero retention after matter close
  • SOC 2 Type II Target
Org

Hosted with org isolation

For mid-size firms and in-house legal ops
  • Annual contract, multi-matter
  • Dedicated org tenant in PlumGen's cloud (logical isolation)
  • SSO (SAML 2.0, OIDC), role-based access, audit export
  • DPA and BAA available
Self-host

In your cloud

For enterprises, regulated industries, air-gapped
  • ORCA deployed into your cloud tenant or private infrastructure
  • Your IAM, your KMS, your network, your logs
  • PlumGen ships software + upgrades; you own the infrastructure
  • Full air-gap and on-prem deployment supported for federal / defense and regulated environments
Comparison

One platform, many vendors retired.

ORCA replaces five categories of stack you may be running today, with one audit trail end to end.

Stack layer you might run todayWhat ORCA covers
Hosted review platformTAR 3.0 + Active Label + Cert Engine
Processing & forensic toolkitPipeline
Collection & web-archival vendorsConnectors
Early case assessment toolsCert Engine + Analysis
Privilege-log spreadsheets / ad-hoc toolsBates + Load Files

ORCA does not replace dedicated legal-hold-tracking or records-management systems. We integrate with them via legal-hold APIs.

Pick the deployment, see the price.

Four tiers, transparent per-matter and annual pricing, no per-GB review fees. See what your matter would cost on each tier — or talk to us for self-host quotes.

See pricing →   Book a walkthrough