Multimodal redaction runtime for sensitive data.
Detect and remove sensitive information across documents, images, and audio. Combines deterministic patterns, NER, vision-language model classification, and whole-audit LLM verification into auditable, policy-driven pipelines built for regulated industries such as healthcare, legal, government, and financial services.
- Multimodal codecs: read, edit, and write PDF, DOCX, images, audio, CSV, JSON, and plain text through a unified span-based content model
- Layered detection: regex, dictionary, and checksum patterns run first at low cost; NER, OCR, VLM, and LLM classification handle what deterministic methods cannot
- Context-aware redaction: mask, replace, hash, encrypt, blur, block, and pixelate with policy-driven rules scoped to entity type, document class, and confidence threshold
- Streaming pipeline: extraction → detection → deduplication → redaction → validation, with concurrent execution across documents and within each phase
The fastest way to get started is with Nvisy Cloud.
For self-hosted deployments, refer to docker/ for compose files and
infrastructure requirements, and Nvisy.example.toml for
the configuration schema.
See docs/ for architecture, security, and API documentation.
See CHANGELOG.md for release notes and version history.
Apache 2.0 License, see LICENSE.txt
- Documentation: docs.nvisy.com
- Issues: GitHub Issues
- Email: support@nvisy.com