Feith is FedRAMP Moderate Authorized → Read the release
Feith
FOIA Workbench

The entire FOIA lifecycle,
on one platform.

Capture, search, review, redact, release, then defend every decision in any audit. AI is built in to keep pace as request volume rises, and the whole platform is FedRAMP authorized.

Book a demo See the lifecycle ›
FOIA Workbench walkthrough 5:01
The urgency

FOIA was never built for this volume.

Federal FOIA programs are underwater, and a growing share of requests now come from AI tools that file faster than any team can process by hand. You can't hire your way out of that. The agencies that keep up will be the ones that put purpose-built AI to work across the FOIA workflow.

Volume up.
Request counts keep growing year over year at most major agencies.
Staff flat.
FOIA office headcount has not kept pace with request volume.
Tools fragmented.
Email, spreadsheets, and case tools that don't talk to each other.
Clock running.
The 20-day statutory clock starts the moment a request arrives.
Audits looming.
OIG and DOJ both expect defensible documentation on every decision.
The platform

From the public portal to the DOJ report.

Ten lifecycle stages. Four systems beneath them. The same case file under every one.

01
Public portal
02
Intake
03
Request management
04
Collection & search
05
Deduplication
06
Document review
07
Redaction
08
AO review & release
09
Appeals & litigation
10
Retention & disposition
Beneath every stage: the platform foundation
Role-based access & permissions
Immutable audit trail
Reporting & analytics
Government-grade security

The lifecycle is one connected system. Every stage shares the same audit trail, the same permission model, and the same data foundation.

Request management

One case file. One source of truth.

Open any request and the full picture is there: tasks, correspondence, deadlines, fee assessment, multi-office coordination, restricted content, and role-based access, all in the same record.

  • 01 Tasks assigned across offices with named owners and deadlines. No chasing updates across email threads.
  • 02 The statutory clock counts down automatically, with tolling controls when the agency needs to pause it for fee-waiver review or request clarification.
  • 03 Requester correspondence captured in the case record. Inbound email routes to the right case; outbound uses response templates and certificate-based e-signatures.
  • 04 Multi-office coordination built in. Configurable workflows route through OIG, Legal, and Records Management, with tasks visible by role, office, and analyst.
  • 05 Role-based access enforced throughout. Restricted content stays restricted; legal-eyes-only material is locked to the legal reviewer pool.

Everything about the request attaches to that one record and shares its single audit trail.

Collection · review · redaction

Find the records. Review them. Redact in the browser.

Collection, deduplication, document review, and redaction all happen in the same system where the request lives. Nothing exports to a separate tool. Nothing imports back in.

Search & collection
  • Pre-built connectors collect responsive records from M365, SharePoint, Documentum, email archives, and file shares straight into the case
  • Bulk upload and drag-and-drop for records that come in manually
  • Email capture pulls responsive correspondence into the case folder automatically
  • OCR runs on scanned documents at ingestion; every record is searchable from arrival
Deduplication
  • Near-duplicate detection removes redundant copies before review begins
  • Exact and near-match identification surfaces version drift across repositories
  • Saves review hours on every case; the larger the case, the bigger the savings
Browser-based redaction
  • Documents and images redacted in-browser with no desktop client required; audio and video sources covered too
  • Exemptions (b)(1) through (b)(9) supported, with a configurable exemption library
  • Bulk redaction for search terms, PII patterns, SSNs, and phone numbers
  • AI-assisted suggestions with human reviewer approval on every action
  • Every redaction logged to the immutable audit trail with user, action, timestamp, and reasoning

Every redaction is a decision. Every decision is logged. Every log entry is defensible.

Reporting · Vaughn index · DOJ Annual Report

The Vaughn index builds itself. So does the DOJ report.

The data that powers FOIA Workbench is the same data that powers FOIA reporting. Because all of it lives in one system, reporting comes straight from the case data you've already captured.

Vaughn index, automatically
  • The document-by-document log of withheld and partially withheld records that defends the agency's exemption decisions in court
  • Built as the case progresses, pulling metadata, exemption codes, and reviewer reasoning entered at redaction time
  • Kept current for every case, so a compliant draft is ready the moment a case moves to appeal or litigation
Reporting that's already running
  • Backlog, processing times, workload by office, exemption usage, denial rates, days-to-close
  • Dashboards are live because the data is live
  • Compile reports at any point in the year; the data is already in the system
DOJ FOIA Annual Report, auto-generated
  • Annual Report draft generated from FOIA Workbench data on requests received, processed, and pending
  • Agency staff review, validate, and edit the draft before submission
  • The platform handles data assembly; the agency handles the editorial

Live data. Live reporting. The DOJ report draft is two clicks away whenever the agency needs it.

After release

The case doesn't end at release.

A FOIA case rarely ends at release. Appeals, litigation, and final disposition all run on the same case file: the same record, correspondence, and Vaughn index that produced the release.

Appeals: same file, new clock
  • The appeals officer picks up the same case file, with the same correspondence history and the same Vaughn index
  • Appeals filed through the public portal link automatically to the original case
  • Tracked as its own statutory process, with its own clock and its own audit trail
Litigation: packaged for export
  • Litigation holds suspend retention timers automatically
  • The full record compiles for counsel: draft Vaughn index, responsive documents, and the exemptions already applied
  • Handed to the litigation team in a ready format, without rework
Retention & disposition: NARA-aligned
  • Closed cases archived to a NARA-aligned repository with retention schedules applied
  • Disposition runs review → approve → transfer / destroy at the end of the retention period
  • Direct integration for agencies running Feith RMA iQ

One file in. Every step after it (appeal, litigation, disposition) runs on the same record, without rework.

Agencies running both FOIA Workbench and Feith RMA iQ get unified records discovery, unified retention, and unified disposition.
Read the RMA iQ overview ›
Beneath everything

Four systems. Every case, every action.

Permissions, audit trail, reporting, and security run platform-wide. None of them is an add-on. All of them are how the platform is built.

Role-based access & permissions

  • Role-based workspaces out of the box: Triage, Action Officer, Reviewer, Manager, Legal, and Billing
  • Custom role definitions for agency-specific organizational structures
  • Need-to-know access controls at the case, document, and field level
  • Redaction authority separation: applying a redaction and approving a redaction are different role permissions
  • Staff authenticate through agency Active Directory and PIV/CAC; public requesters sign in to the portal with Login.gov or ID.me

Immutable audit trail

  • Every action recorded: who, what, when, and why
  • Exportable as an evidence package for OIG audits, FOIA appeals, and litigation
  • Log shipping to the agency's SIEM, with alerts on anomalous access
  • Computed hash digests keep the audit trail tamper-evident: any alteration of a record is immediately detectable
  • Retention of audit logs governed by agency policy and applied automatically

Reporting & analytics

  • Real-time dashboards for backlog, processing time, workload, exemption usage, fee assessment
  • Custom reports through an embedded query builder: filters and pivot tables, no query writing
  • DOJ FOIA Annual Report auto-generated from platform data
  • Configurable KPIs for agency-specific reporting needs
  • Drill-down from agency-level to office-level to individual analyst views

Government-grade security

  • NIST 800-53 control framework applied platform-wide
  • FIPS-validated cryptography for data at rest and in transit
  • Every deployment is a dedicated, isolated single-tenant instance — never shared infrastructure
  • Delivered in a FedRAMP-authorized cloud — the full compliance list is under Security & compliance below
Audit-grade AI

AI for the bottleneck. Reasoning at every step.

For most agencies, redaction is the largest FOIA bottleneck, and the place where AI helps most, when it's configured correctly and reviewed by humans who stay in charge of every decision.

01
Prompt Runner: the AI step

FOIA Workbench runs its AI steps through Prompt Runner, Feith's AI step for bringing language models into agency workflows. A language model never decides what to do next; agency policy is the workflow. Each step is deterministic: AI performs bounded tasks under strict schemas, with validation at every checkpoint.

02
Grounded AI redaction

By the time responsive records reach the AO, the AI redaction suite has already reviewed the documents, flagged potentially exempt content, and suggested redactions grounded in FOIA law. Each suggestion carries an exemption code, (b)(1) through (b)(9), and plain-language reasoning. Unlike traditional ML that needs months of agency-specific training, this model ships ready out of the box, because it's built on the law itself.

03
More AI capabilities through Prompt Runner
  • Summarization of long responsive records for AO review
  • Classification of incoming requests for routing and triage
  • Metadata extraction (dates, requester identities, exemption criteria)
  • Drafting of acknowledgment letters and standard responses
  • Relevance review on bulk record collections
  • Search-term and Boolean query suggestions during collection
Deterministic AI workflow Prompt Runner
Agency policy
Workflow root
Bounded AI task
schema-validated input
Bounded AI task
schema-validated input
Bounded AI task
schema-validated input
Human reviewer approval
Bounded by design
?
Autonomous agent
open-ended planning · behavioral drift
See also
AI in FOIA Workbench is one application of Feith's broader AI capability set. Read the AI in Feith overview ›
Building agents?
Feith MCP is platform-wide, so FOIA cases, documents, audit, and reports are reachable through Model Context Protocol like any other Feith surface. See Feith MCP ›
Security & compliance

Built for the standards your agency answers to.

FOIA Workbench is delivered in a FedRAMP Moderate-authorized cloud environment, and meets the security and records-management standards federal agencies are required to meet.

FedRAMP Moderate
Platform security
Authorized cloud environment
Active ATO
Platform security
In production at federal agencies
DISA-JITC tested
Platform security
Legacy DoD 5015.02-STD
FIPS 140-2 / 140-3
Platform security
Validated cryptographic modules
NIST 800-53
Platform security
Security control framework
NARA 36 CFR 1236
Records management
Electronic records management
Federal Records Act
Records management
44 U.S.C. Chapter 31
OMB M-19-21
Records management
Electronic records mandate
Tamper-evident archival
Records management
Secure hash verification
Retention & disposition
Records management
Applied at case closure
SAML 2.0 · MFA
Authentication & access
Multifactor single sign-on
RBAC
Authentication & access
Case, document & field level
ABAC
Authentication & access
Record- & field-level attribute rules
PIV / CAC
Authentication & access
Smart-card auth for federal staff
LDAP / Active Directory
Authentication & access
Directory integration
What the work looks like

The numbers behind defensible FOIA.

20 days
Statutory clock, tracked automatically
Tolling controls for fee review and request perfection
9
FOIA exemptions supported, (b)(1) through (b)(9)
Plus CUI, PII, and agency-configurable exemption sets
1
Case file per request
Tasks, correspondence, redaction, and audit, all in one record
0
Custom code required to deploy
Low-code workflow configuration, ships ready for federal use
State, local & enterprise
The same platform, configured for your records law.
State & local
CPRA · OPRA · GIPA · state PRAs
Higher education
FERPA · state public records
Healthcare
HIPAA · state health records law
Regulated commercial
Regulator records requests · state PRAs
Talk to us
Get started

See FOIA Workbench on your records.

Bring a sample of your FOIA queue, your exemption library, and a redacted case from last quarter. We run a walkthrough on your actual workflow and show you what the audit trail produces.

Schedule a walkthrough