FLASH Intelligence
Local
🔍

Private RAG

Ingest documents with client-side encryption, then retrieve by semantic similarity.

Open RAG →
🧠

Agent Memory

Encrypted episodic memory for AI agents — recall facts without cloud retention.

Open Memory →
🔐

Sealed Vault

Passphrase-isolated secrets with auto-lock — separate from general collections.

Open Vault →
Server-blindPlaintext never reaches storage or query engine
Local-firstYour key, your disk, your rules
Intelligence-nativeRAG + memory + vault in one engine

Collection

Target knowledge base name

Ingest

Encrypt & index new knowledge

Response
Waiting for ingest…

Semantic Search

Retrieve context for LLM pipelines

Context pack
Results appear here…

Remember

Store a new memory

Last action
—

Recall

Search by meaning

Access

Unlock to manage records

Locked

Integrity Proof

Signed Merkle manifest

Manifest
—

Prompt Firewall

PII & secret detection

Result
—

Collections

Quick insert

Documents

Select a collection