Company-developed products
From a working product to an agency’s requirement
Shared building blocks run through all of them: document intelligence with every extracted value tied to its source, AI that a person can confirm or correct, offline and on-premises operation, role-based access, audit trails, and accessible, multilingual interfaces.
GovAccess
Digital accessibility operations for public-sector websites and documents
In production
Scans a public website, scores what it finds, and turns automated accessibility findings into a prioritized plan. Repairs and tags PDF documents, produces accessible HTML alternate versions, and tracks the work to completion.
What is built
- Site crawler with automated accessibility testing and a deterministic readiness score
- PDF repair and tagging, with OCR for scanned documents
- AI-written image descriptions and accessible HTML alternates — switchable off per organization
- Multi-tenant organizations, roles, and document inventory
How it can serve an agency
Built for state and local governments and public universities preparing for the ADA Title II web and document deadlines. The document-repair pipeline can be run on agency infrastructure for document backlogs that cannot leave it.
Capabilities
- Accessibility
- Document processing
- OCR
- AI
GovEnclave
Offline contract intelligence that runs on the agency’s own server
Pre-pilot
Reads contracts and pulls out the terms staff track by hand — expiration dates, renewals, notice periods, insurance requirements — using a language model that runs entirely on local hardware, with no data sent to an outside AI service.
What is built
- PDF, Word and text ingestion with OCR fallback and per-page coverage reporting
- Local language model with schema-constrained output; every value cites its page and passage
- Review queue to confirm or correct each extracted value
- Role-based access per collection and a hash-chained, append-only audit log
How it can serve an agency
For counties, cities and special districts whose contracts cannot be sent to a public AI service. The same offline, audited pipeline extends to insurance certificates, invoices and grant documents.
Capabilities
- Document intelligence
- Local AI
- Audit trail
- Offline
SH Help
AI assistant that connects people to verified community resources
Release candidate
A mobile-first assistant that helps a person describe a need and find the right local resource. Answers come only from a curated, reviewed resource database; conversations are not stored, and a person decides what summary, if any, is passed to a helper.
What is built
- Retrieval-grounded answers that cannot invent a phone number or an eligibility rule
- Deterministic crisis-escalation layer that runs before any model call
- English, Spanish and Portuguese; installable on a phone
- Admin review of sources, role-based access, audit log and retention controls
How it can serve an agency
The shape of a benefits and resource navigator for a human-services agency: grounded answers from the agency’s own program directory, caseworker routing, and no transcript retention.
Capabilities
- AI assistant
- RAG
- Multilingual
- Privacy
MissionLayer Sizing
Contactless body and foot measurement from phone photos
Prototype
Measures height, body girths and foot dimensions from photos taken against printed reference targets, then recommends garment and footwear sizes. Runs offline, on a local server, with no contact and no specialist scanner.
What is built
- Computer vision with printed reference targets, plane fitting and quality checks
- On-device segmentation and pose models — no cloud processing
- Encrypted storage, role tokens, and a hash-chained audit log
- Accessibility-tested capture interface
How it can serve an agency
Uniform, boot and protective-equipment sizing for military, veterans and first-responder programs, including remote or mail-in fitting. Accuracy is measured on synthetic data so far; a field validation study is designed and not yet run.
Capabilities
- Computer vision
- On-device ML
- Offline
PdayPost
Turns a long run of letters and photos into a searchable, sourced archive
Staging
Receives weekly letters and photos by email and organizes them into a timeline, a map and a printed book. AI reads each letter and records the facts in it, and every fact links back to the passage it came from.
What is built
- Email intake with malware scanning of attachments
- AI fact extraction with per-fact provenance and a human confirm-or-correct history
- Allow-listed, version-pinned AI models evaluated against a reference set in CI
- English, Spanish and Portuguese; WCAG 2.2 AA audit; personal data sealed and deleted on schedule
How it can serve an agency
The same intake and provenance pipeline fits correspondence and records work: constituent-letter triage, records intake, and request processing where every extracted fact must point to its source.
Capabilities
- Document intelligence
- AI
- Provenance
- Multilingual
MeetingNotes
Records a meeting on a phone and produces minutes with assigned tasks
Demonstration
Records meetings of up to an hour on a phone, even with poor connectivity, transcribes them, and drafts minutes with each action item assigned to a participant. Built as a working demonstration for a request for proposals.
What is built
- Offline-safe recording in short segments, with a resilient upload queue
- Speech-to-text transcription and AI-drafted minutes in a fixed structure
- Never invents an owner or a date — unclear items are marked undetermined
- Editable minutes, PDF export, and a per-participant task email
How it can serve an agency
Minutes for boards and public meetings, field inspections and outreach interviews in low-connectivity areas. An agency deployment would add sign-in, audit logging, records retention and approved speech and AI endpoints.
Capabilities
- Speech-to-text
- AI
- Offline
- Mobile
Rejoice
Join-by-QR participation app with a live, privacy-preserving dashboard
Release candidate
People join a community challenge by scanning a QR code, without giving a name or email, and log a daily check-in on a private calendar. A projector page shows live totals, and group figures are hidden until a group is large enough to stay anonymous.
What is built
- Anonymous sign-up in under a minute, protected by a CAPTCHA
- Realtime aggregate counters for a live display
- Small-group suppression so no individual can be identified
- Row-level security throughout; load-tested for hundreds of sign-ups in minutes
How it can serve an agency
Public-health and civic campaigns, event check-in and volunteer pledges where an agency wants live participation numbers without collecting personal information.
Capabilities
- Privacy by design
- Realtime
- Mobile
FiSa Invest
Multi-model AI research engine with a public, timestamped record
Release candidate
Several independent AI models analyze the same market data, and a finding is published only when they agree; disagreement is recorded rather than hidden. Includes AI summaries of public financial filings and a timestamped record of every output.
What is built
- Consensus and divergence across independent AI models from different providers
- AI summaries of public filings
- Timestamped, auditable record of every published output and its cost
- Read-only account linking with encrypted tokens; row-level security
How it can serve an agency
The pattern for high-stakes analysis where one model is not enough: independent models must agree, disagreement is surfaced for a person, and every output is attributable and auditable — for example, review of public filings or grant applications.
Capabilities
- Multi-model AI
- Auditability
- Document summarization
Products designed and built by MissionLayer Systems LLC. They demonstrate capability and can be shown working; they are not contract past performance, and none is FedRAMP authorized or certifies compliance with any standard. Status as of October 2026.