nvisy.com

The self-hosted backend service for the platform, exposing REST and WebSocket interfaces to support document processing, external integrations, and data management. Requires PostgreSQL, S3 and NATS.

Composable redaction framework for building PII detection and removal pipelines over unstructured data. Built-in recognizers, custom rules, and ML models to analyze and anonymize multimodal documents.

Web and desktop app for detecting and redacting PII in documents, built with Nuxt and Tauri. Upload files, configure detection rules, and process everything locally in the browser or a native client.

Official JavaScript and TypeScript SDK for the platform, offering a modern, type-safe client with flexible configuration and seamless API integration.

OCR, NER, and vision-language inference services implementing Elide's recognizer contracts, self-hosted for PII detection. Ships as BentoML packages and a client crate with NER, OCR, and STT backends.

Synthetic document corpora with known ground truth, for benchmarking redaction accuracy. Generates fabricated PII planted at recorded positions, then scores recall, precision, and boundary accuracy per entity type and modality.

Stateless document pipeline built on the Elide toolkit, providing wire schemas, policy-driven redaction governance, compliance templates for common privacy regimes, and the runtime engine that ties them together.

Official Model Context Protocol server for the platform, exposing redaction pipelines, policies, and detections as tools for AI agents.

Official documentation for Nvisy. Setup and self-hosting guides, SDK usage, redaction-policy reference, and a full OpenAPI spec. Built with Mintlify.

Official Rust SDK for the platform, delivering a performant, type-safe client with idiomatic Rust APIs, flexible configuration, and seamless integration with the platform’s services.