Compute & infrastructure
Model inference, training experiments, GPU access, cloud servers, storage, vector databases, APIs, domains, and deployment. These are ongoing monthly costs.
Data work
Cleaning, labeling, verifying, embedding, hosting, and documenting datasets — including 22,000+ historical records. This is labor and storage, not a one-time task.
Equipment
GPUs, workstation upgrades, storage drives, sensors, networking hardware, and test components needed to build and validate instruments.
Development time
Uninterrupted time to write code, test systems, document results, and publish evidence — the work that has to happen before there is a product to sell.
Legal & operational basics
Forming the entity, licensing review, data security, and hosting. These costs land before any revenue exists.
Open, free output
Tools, datasets, research notes, benchmarks, and public demonstrations are published for anyone to use. Support keeps them free.
It funds work before revenue
Useful experiments, prototypes, and benchmarks have to be built before anything is ready to sell. Contributions cover that gap between personal funding and future grants, contracts, or licensing.
It preserves independence
Support lets the lab do exploratory and public-benefit research — privacy-first systems, local-first models, transparent reasoning, archival preservation — without giving up ownership, control, or the mission to outside investors.
It is a signal, not a purchase
A contribution says the research is worth continuing. You are not buying a product or making an investment — you are enabling a direction you want to see developed.