gharasathi
A fully local home AI system that integrates household finance, photos and documents. It emphasizes data privacy and verifiable reasoning, and runs on a Mac Studio.
Timeline (last 90 days)
gharasathi is an AI system that runs entirely inside the home, integrating money, photos and documents, with inference running on a single Mac Studio and data never sent to external clouds.
gharasathi follows three design principles: data stays at home, code does the computation while the model does the explanation, and every finding shows its evidence.
The gharasathi project started out on a $200 mini PC, centered on a Neo4j graph database and paired with a small 3B-parameter model.
Hishob is gharasathi's finance module, a private money dashboard that reads bank data, bills and statements, finds where the money goes and explains what to do about it.
gharasathi currently runs Qwen3.5-122B on the Mac Studio as the household financial analyst.
One article in the gharasathi series, "Reading Bills and Statements Without AI, and the Money They Found", discusses reading bills and statements without using AI.
gharasathi's approach puts deterministic parsing first, with the model only responsible for explaining the structured results that parsing produces.
gharasathi's roadmap includes deploying Immich on the Mac Studio to move the family photo library off iCloud and Google Photos.
Two articles in the gharasathi series, "Running an LLM on 6GB RAM: Model Selection for Edge AI" and "What OpenClaw Teaches Us About Personal AI Security", discuss edge AI model selection and personal AI security boundaries.
In household finance scenarios, mandatorily showing the raw evidence (such as transaction records or the original document text) is key to building user trust.
All Facts (10)
In household finance scenarios, mandatorily showing the raw evidence (such as transaction records or the original document text) is key to building user trust.
50%Unverifiedgharasathi follows three design principles: data stays at home, code does the computation while the model does the explanation, and every finding shows its evidence.
50%UnverifiedThe gharasathi project started out on a $200 mini PC, centered on a Neo4j graph database and paired with a small 3B-parameter model.
50%UnverifiedHishob is gharasathi's finance module, a private money dashboard that reads bank data, bills and statements, finds where the money goes and explains what to do about it.
50%Unverifiedgharasathi currently runs Qwen3.5-122B on the Mac Studio as the household financial analyst.
50%UnverifiedOne article in the gharasathi series, "Reading Bills and Statements Without AI, and the Money They Found", discusses reading bills and statements without using AI.
50%Unverifiedgharasathi's approach puts deterministic parsing first, with the model only responsible for explaining the structured results that parsing produces.
50%Unverifiedgharasathi is an AI system that runs entirely inside the home, integrating money, photos and documents, with inference running on a single Mac Studio and data never sent to external clouds.
50%UnverifiedTwo articles in the gharasathi series, "Running an LLM on 6GB RAM: Model Selection for Edge AI" and "What OpenClaw Teaches Us About Personal AI Security", discuss edge AI model selection and personal AI security boundaries.
50%Unverifiedgharasathi's roadmap includes deploying Immich on the Mac Studio to move the family photo library off iCloud and Google Photos.
50%