AegisTextPrivacy infrastructure for production text pipelines

A precision redaction layer for teams that need to remove sensitive data before it reaches LLMs, RAG pipelines, search indexes, or classifiers.

Redaction-first API

Use POST /v1/redact to convert raw text into privacy-safe redacted_text without redesigning your downstream stack.

Insert AegisText between raw text and downstream AI systems without changing the rest of your pipeline.

Keep text readable, preserve structure, and opt into placeholder_map only when your workflow truly requires reversibility.

Public docs, examples, payload fixtures, and contract tests are aligned to a single /v1/redact model.

http

POST /v1/redact

Use /v1/redact to transform raw text into redacted_text before sending it downstream.

Teams sending customer text to external LLMs
Products building RAG on internal documents or support logs
Engineering teams that need privacy-safe prompts without redesigning existing systems
Organizations that need a practical redaction layer before search, indexing, or classification

Send support notes to /v1/redact, then forward redacted_text to an LLM summarizer.

Redact sensitive spans before indexing text into a vector store or search layer.

Apply placeholder-based redaction before prompts reach downstream models.