# Rigr AI

> Secure AI triage for high-stakes evidence and safety workflows. Collaborative investigation platform for law enforcement, plus product-grade apparent-age estimation for Trust & Safety teams. AI-powered video, image, audio, and text analysis — deployed on-premise, air-gapped, or in your cloud. Ireland-based.

Rigr AI builds a platform called **VST Teams** (Video / Still / Text), currently at **release v1.11.57**, and a set of enrichment capabilities used by national and supranational law-enforcement agencies, and by EU-funded defence research consortia. VST Teams is a full collaborative investigation platform: multi-user workflows with named sessions, per-user bookmarks and annotations, automatic entity extraction (Named Entity Recognition for people, organisations, locations, and dates), unified image + video gallery, and folder-level triage. Since v1.6 (April 2026) it has added known-material hash matching against shared law-enforcement lists, a dedicated classification review workflow, ingestion of email and document evidence, and search across transcribed speech, on-screen text and named faces.

Rigr AI also serves Trust & Safety teams directly: the Age Estimation capability returns a calibrated apparent-age estimate with uncertainty, letting platforms route age-ambiguous AI-generated and user-uploaded content to the right review queue.

Everything Rigr ships is designed for regulated environments: fully containerised, deployable on-premise, air-gapped, or in your agency's cloud, no external data dependency in air-gapped mode, no data retention, customer retains full control of inputs and outputs. Designed to support GDPR, the Law Enforcement Directive (EU 2016/680), and the EU AI Act (2021/0106).

Public content below is safe to crawl and cite. Annex content (technical specifications for police and defence audiences) is gated behind vetted-access authentication and is intentionally not linked here.

## Platform

- [VST Teams platform overview](https://rigr.ai/platform): collaborative investigation platform for synchronised multi-view incident review (align every camera on one incident — bodycam, dashcam, CCTV, bystander footage — onto one timeline from shared audio), video navigation, transcription & speaker identification, face detection, text extraction (OCR), entity extraction (NER), multi-user workflows, and folder-level triage. Deployed on-premise, air-gapped, or cloud.
- [VST Teams Capability Brief, Police Edition (PDF)](https://rigr.ai/resources/vst-teams-capability-brief-police.pdf): six-page A4 brief for police and regulated investigative teams — core capabilities, enrichment modules, deployment and information control, where Rigr sits in the forensic workflow, and media classification. Its full text is reproduced in llms-full.txt, since a PDF is not readable by an agent.
- [VST Teams v1.6 brochure (PDF)](https://rigr.ai/resources/vst-teams-v1.6.pdf): the earlier printed feature list, kept at its original URL; superseded by the capability brief above.
- [Deployment, sovereignty & information control](https://rigr.ai/deployment): how Rigr deploys into regulated environments — on-premise, air-gapped, private sovereign AI, secure private cloud, and government cloud.
- [Why Rigr AI](https://rigr.ai/why-rigr-ai): design principles behind the platform.
- [Use cases](https://rigr.ai/use-cases): operational scenarios.

## Solutions

- [Digital Evidence Backlogs](https://rigr.ai/solutions/digital-evidence-backlog): buyer-led page for police and regulated investigative teams — reducing digital evidence backlogs without compromising information control.
- [Criminal Defense Video Evidence](https://rigr.ai/solutions/criminal-defense-video-evidence): buyer-led page for US criminal-defense practices — synchronise every camera on an incident (bodycam, dashcam, CCTV, a bystander's phone, from any vendor) onto one timeline and watch every angle at once, plus a searchable, translated, per-camera transcript, on a workstation your organisation controls.
- [Trust & Safety Image Moderation](https://rigr.ai/solutions/trust-safety-age-estimation): buyer-led page for Trust & Safety and platform teams — image moderation and apparent-age estimation from one API call: nudity, exposed intimate anatomy and sexual-activity detection alongside apparent age and developmental stage, with calibrated uncertainty and configurable review bands for routing user uploads.

## Capabilities

- [Age Estimation](https://rigr.ai/capabilities/age-estimation): AI-based age estimation from facial imagery, returning a point estimate and calibrated uncertainty; used in sensitive law-enforcement investigations involving minors and by Trust & Safety teams for content review routing.
- [Age Estimation API reference](https://rigr.ai/docs/age-estimation-api): REST reference for the facial age estimation API — endpoints, request/response schemas, error codes, image limits, and self-hosted Docker deployment.
- [Media Examiner API reference](https://rigr.ai/docs/media-examiner-api): REST reference for the integrated image-analysis API — nudity, intimate-anatomy and sexual-content detection with apparent-age estimation in one call. Endpoints, response schema, severity scale, contextual flags, limits and error codes. Machine-readable OpenAPI 3.1 document at https://rigr.ai/openapi/media-examiner.yaml.
- [Image Geolocation — Morrigan](https://rigr.ai/capabilities/location-estimation): AI image geolocation / visual geolocation from images and video, indoors and outdoors, ranked with confidence scores.
- [CSAM Classification](https://rigr.ai/capabilities/media-classification): AI CSAM detection and severity classification for investigations and moderation at scale, with contextual flags and a reviewer verdict alongside the machine call. Interactive tour under "Product tours" below.
- [Transcription & Summarisation](https://rigr.ai/capabilities/summarisation): transcription, translation and summarisation of interviews, calls, chat logs and multilingual text evidence, with every summary tied back to its source passage.

## Product tours

Canned, click-through tours of the real VST Teams screens. Sample data only, no sign-in, nothing is processed — safe for agents and people to explore and cite.

- [Product tours index](https://rigr.ai/tours/): hub for all Rigr AI product tours.
- [Media Classification — the review tour](https://rigr.ai/tours/classification-review/): how a team triages and reviews seized media at scale — starting with the worst rather than the first, cutting a large case down to what needs a person, grading severity on apparent age plus the specific act, and treating a known-material hash match as fact rather than opinion. Available in all eight site languages under the matching `/{lang}/tours/classification-review/` path.
- [Bodycam review — the criminal-defense workflow](https://rigr.ai/tours/bodycam-review/): how a criminal-defense team turns bodycam, a bystander phone, CCTV and in-car audio into one searchable, translated timeline — audio-synced across cameras onto a single incident clock, honest about the cameras it cannot confidently place (named and set aside, never faked onto the clock), with a merged transcript that surfaces contradictions across feeds, Original/Translated views for non-English audio, and a court-defensible package export behind a redaction-review gate. English only.
- [Redaction & Disclosure — the lifecycle](https://rigr.ai/tours/redaction-review/): how a team takes a seizure from intake to a disclosable copy — VST proposes every redaction candidate, a reviewer decides each, and one decision fans across documents, transcripts and on-screen text, gated behind a complete, audited review before anything is disclosed. Every frame is synthetic. English only.
- [Location — Morrigan](https://rigr.ai/tours/location-review/): how an investigator estimates where an image or video was taken — indoors or outdoors — from the scene itself, ranked with confidence and shown alongside the visual evidence the model is reading. Sample data only. English only.

## Projects

- [ARICA](https://rigr.ai/projects/arica): EU/EACTDA/Home Office-funded project on Assessing Risk Indicators of Child Abuse — Rigr contributes core AI enrichment for text, images, and video. Public site: https://aricaproject.eu.
- [BadB](https://rigr.ai/projects/badb): EU-funded defence research on GNSS-denied navigation for uncrewed vehicles — Rigr contributes software, simulation, and ML. Public site: https://badbproject.eu.

## Articles

- [Articles](https://rigr.ai/articles/): Edward Dixon's occasional writing on AI, investigation tooling, defence research, and the systems Rigr builds. Low cadence, long-form, not promotional. RSS at https://rigr.ai/articles/rss.xml.

## Interactive demos

- [Air Defence Allocation Sim](https://rigr.ai/sim/): browser-based, deterministic simulation of the trade-off between fixed point air defence, mobile combat air patrol, sensor warning, and saturation strike. Defenders allocate budget, attackers vary concentration and wave structure; outputs preserved-value, intercept geometry, magazine utilisation, plus a written per-run analysis with a same-seed flipped-allocation counterfactual. Runs in any modern browser. Abstract — no real systems, weapons, or countries are modelled.
- [Convoy Protection Overwatch Sim](https://rigr.ai/sim/convoy/): browser-based, deterministic simulation of defending a small armoured patrol against drone ambush from close terrain. 2.5D terrain line-of-sight — polygonal forest blocks with canopy height ray-checked against sensor heights (hull, mast, overwatch) and drone altitude — with the resulting dead ground shaded live on the map. Models hard-kill reaction chains, fire-channel saturation, and sensor-only vs armed top cover; every run is paired with a same-seed armed-overwatch counterfactual and a written analysis. Abstract — no real systems or weapons are modelled.

## Contact

- [Request a briefing](https://rigr.ai/contact): form for prospective customers, partners, and press.
- [Get a free API key](https://rigr.ai/get-api-key): self-serve signup for a free Age Estimation API key, for individual developers and investigators evaluating the API directly (as opposed to VST Teams or a private pilot).

## Languages

Public pages are available in English (default, no prefix), French (`/fr/`), German (`/de/`), Spanish (`/es/`), Portuguese (`/pt/`), Italian (`/it/`), Traditional Chinese for Hong Kong (`/zh/`, served with `lang="zh-HK"`), and Japanese (`/ja/`). URLs and slugs are in English across all languages.

## Not for crawling

The following paths are gated and disallowed in [robots.txt](https://rigr.ai/robots.txt):

- `/admin/` — back-office dashboard
- `/access/` — authentication flows for vetted annex access
- `/decks/` — presentation decks served only after server-side auth
- `/api/annex/` — protected annex content
- `/prototypes/` — internal UI previews, shareable by URL but not advertised or indexed

Please respect these boundaries. A concise expanded view of public content is at [llms-full.txt](https://rigr.ai/llms-full.txt).