Turn any document into accurate, private AI search.

troveGEN is a RAG pipeline and search API built by Intellara Technologies. It turns documents into accurate, private AI search: it reads them, protects sensitive data, indexes them, and returns the right passages with sources to your app or AI agent.

Two groups. AI app and agent builders use it to get the right context to generate from through one API. Engineering teams and regulated organisations use it to keep fast-growing document sets — specs, runbooks, contracts, statements, records — searchable with permissions and versioning.

It reads documents (with OCR for scans), protects sensitive values, splits content on its natural structure, indexes it for both keyword and semantic search, and then answers queries by fusing and reranking results, checking permissions inside the query, and returning cited passages.

PDFs and scans, Word, PowerPoint, spreadsheets, web pages and feeds, and email (.eml). Eighteen languages are verified end to end. Chinese, Japanese, Korean and Thai are not supported yet.

In troveGEN-Z, troveGEN hosts vectors, models and OCR in a dedicated schema and namespace for you. In bring-your-own mode your vectors stay in your own Qdrant, Pinecone, pgvector or Postgres. In Sovereign mode everything can run air-gapped inside your environment.

Account numbers, national IDs and similar values are replaced with tokens before indexing, with originals in a separate encrypted vault. Permissions are applied inside every search using signed end-user assertions and identity-provider group membership, and the filter fails closed.

Retrieval is the core of the product. Cited answer generation is optional: citations are checked against the real passages, unanswerable questions are refused, and an agentic mode can plan several searches within a fixed budget.

Through a REST API, a typed TypeScript SDK (@intellara/trovegen-sdk) and an MCP server (@intellara/trovegen-mcp) for AI agents, which is read-only unless writes are explicitly allowed. Connectors keep S3, Azure Blob, Google Drive and web sources in sync.

There is a free trial of 500 pages. Managed troveGEN-Z plans range from $25 to $1,999 per month, bring-your-own-vector-database plans are $89 and $399 per month, and Sovereign deployments are quoted. Usage beyond a plan's limits is refused with a clear message rather than silently metered.

Not yet. No SOC 2 or ISO 27001 audit has been completed. troveGEN relies on architecture — tokenised PII, permission-aware retrieval, per-customer isolation, deletion that propagates, and the option to keep everything in your own environment.