100% on-device. Zero cloud APIs.

AI that never sees your data leave your device.

Every competitor sends your transactions to cloud servers for AI processing. PocketVault runs everything locally.

Local LLM inference

PocketVault integrates llama.cpp for on-device large language model inference. Choose from multiple model tiers (fast, balanced, powerful) based on your device's capabilities.

  • Choose from multiple model tiers
  • Adapts to your device's capabilities
  • No internet required for AI features

LLM Engine

Runtime llama.cpp
Models Fast | Balanced | Powerful
Processing 100% on-device
No data leaves your device
"coffee purchases" AI

Semantic matches

Starbucks 98%
Blue Bottle 95%
Dunkin' 91%

Semantic vector search

Search transactions by meaning using all-MiniLM-L6-v2 ONNX embeddings. Search "coffee purchases" and find Starbucks, Dunkin', and local cafes. Hybrid search combines semantic + keyword matching.

Semantic search uses on-device ONNX embeddings. Results are ranked by meaning similarity, not just keyword matches.

Smart categorization

The AI learns from your patterns and suggests categories for new transactions. Personalized to your spending, improving as you use it.

Learns

Adapts from your spending patterns

Suggests

Auto-categorizes new entries

Improves

Gets smarter over time

Natural language queries

Ask "How much did I spend on food last month?" and get streaming answers powered by on-device RAG that pulls context from your transaction history. The AI can also take actions — create budgets, tag transactions, or generate reports through tool-calling.

How much did I spend on food last month?
$342.50 across 23 transactions. That's 15% of your total spending.
Powered by on-device RAG

Take control of your financial privacy.

Try the finance app that respects your data. Free to start. Premium from $1/month. No account required.