Security
Your data is your data.
You handle privileged communications, trade secrets, and sensitive personal information every day. We built Aquiles with the assumption that every byte of your data is confidential. Because it is.
Bank-Grade Encryption
All data at rest is encrypted using AES-256, the same standard used by financial institutions and government agencies. Credentials are additionally protected with OS-level secure storage.
Local-First Architecture
Your documents, case logs, and search indexes live on your machine — not on our servers, not in a shared cloud. Aquiles works offline. Your data never leaves your device unless you explicitly request an AI analysis.
Zero-Retention Processing
When AI analysis requires cloud processing, only the minimum necessary context is sent — processed in real-time with zero retention. Your data is never stored on third-party servers and never used to train AI models.
Technical details
How it works
Data at rest
All case data — documents, case logs, workspace metadata — is stored in an encrypted local database on your machine. Vector embeddings for semantic search are generated locally and stored locally. There is no server, no cloud sync, and no telemetry on your case content.
Data in transit
The only time data leaves your machine is when you use AI features that require cloud processing. Aquiles sends only the specific context needed for your query — a relevant document excerpt, your question, and minimal workspace metadata. This is transmitted over TLS 1.3 encrypted connections. No data is retained, logged, or stored after processing.
AI model training
Your data is never used to train AI models. We use API-level access to AI providers with explicit zero-training agreements. The providers we work with contractually guarantee that API inputs are not used for model training, fine-tuning, or any form of model improvement. Your case data in, your answer out, nothing retained.
Credential security
API keys and credentials are encrypted using your operating system's native secure storage. These credentials are never stored in plain text, never written to config files, and never transmitted except to their intended API endpoint.
Semantic search indexing
When Aquiles indexes your documents for semantic search, the embedding generation happens entirely on your machine using a lightweight local model. No document content is sent to any external service for indexing.
Scorecard AI validation
AI responses undergo a multi-pass scorecard validation before reaching you. A second analysis pass reviews each response against your source documents — checking for unsupported claims, verifying cited passages actually exist, flagging logical inconsistencies, and scoring accuracy and completeness. This doesn't eliminate the need for professional judgment, but it materially reduces the risk of hallucinated content, fabricated citations, and unsupported assertions making it into your work product.
Our commitment
- We will never sell, share, or monetize your case data.
- We will never use your data to train or improve AI models.
- We will never store your case data on our servers.
- We will always give you full visibility into what data is transmitted and why.
- We will always let you use Aquiles fully offline, without any cloud features.
Questions about our security practices? Contact us at security@aquiles.legal.