Solutions
Froglet
Share data and tools. On your terms.
_o..o_ froglet
Useful work deserves to be used.
Share data and tools. On your terms.
Let someone look up information or run a useful task. Choose what they can use, review it with your AI assistant, and share a link.
Public beta. Try it in your browser without signup or installation.
What is Froglet?
Froglet: share something people can actually use.
Froglet is open-source software, released under the Apache-2.0 licence, for making data and useful tools available to other people and their AI assistants. A file gives someone a copy. A Froglet service lets them ask a question, run a task, or train across organisations and get a verifiable result.
01 / Data
A useful collection
A research catalog, a product list, or a table that translates your terminology into someone else’s. Share only the fields you choose.
02 / Functions
A repeatable task
A calculation, a lookup or a small analysis. Someone provides an input; your tool returns the answer.
03 / Federated learning
Cross-silo compute
Let partner organisations run bounded training steps or map-reduce queries over your local cohort without moving raw records.
How sharing works
How Froglet works: a simple way to share it.
Bob has something useful. Alice needs it. Three steps connect them, and nothing is published until you approve it.
01 / Choose
Pick what to share.
Ask your assistant to prepare a useful part of your work. Only the selected data goes into the shared copy.
02 / Review
See it work first.
See what is included, what is left out, and an example result. You approve before anything is published.
03 / Share
Send a link.
The recipient’s assistant reads what the service does and helps decide whether to use it. Invitation access is sent separately.
Federated learning & distributed compute · Beta
Froglet federated learning: train across organisations. Keep raw data in each silo.
Hospitals, banks and labs can train together without opening inbound firewall ports or exporting sensitive rows.
- Flower (flwr)
- NVIDIA FLARE
- PSI + SplitNN
- OMOP CDM v5.4
- HL7 FHIR R4
- Financial AML
- SecAgg
- DP-SGD
- Wasm
- Beta
Zero raw data movement
Local by design
Each organisation runs its own Froglet node on local data. Updates use pairwise secure-aggregation masking and differential privacy.
Locked compute
Same code everywhere
Every silo pins the same WebAssembly binary and enforces a persistent privacy-budget ledger.
Audit & settlement
Signed receipts
Every training round produces a signed receipt that can be verified offline, with Byzantine-robust aggregation.
Step 01 / Cohort & PSI prep
Map-reduce & vertical alignment
Check cohort counts, compute global normalisation across OMOP, FHIR or AML schemas, or intersect entity IDs through blinded HMAC-SHA256 PSI with zero raw rows exposed.
Step 02 / Local SGD & SplitNN
Flower, NVIDIA FLARE & LoRA
Dispatch Logistic, MLP, LoRA or vertical SplitNN steps through turnkey adapters, with pairwise SecAgg masking and privacy-budget enforcement.
Step 03 / Audit & settlement
Multi-Krum & Shapley lineage
Verify every receipt chain offline, reject Byzantine gradients with MultiKrum, and gate Lightning settlement on a positive Shapley value.
You decide the scope
Share the useful part. Keep the rest yours.
- Choose the data and the task. Review the exact fields or operation others can use.
- Decide who can use it. Start public, or set up invitations with expiry and request limits.
- Change your mind. Review updates before publishing, pause the service, or revoke invitations to stop new work.
A shared result can be kept by its recipient: stopping access prevents new work but cannot recall a result already delivered. Froglet is in public beta and runs on a computer you control.
For organizations
Collaboration shouldn’t start with an infrastructure project. Agree what can be shared, who can use it, and under which conditions.
What you can control today
Our direction is to turn those decisions into a working service. Today Froglet provides selected data sharing, cross-silo federated learning, invitations, execution limits and signed records. Broader contract and SLA automation is still ahead.
Independent workspaces, connected directly, without a central platform.
A few practical questions
Understand it before you start.
Is this like a shared folder?
That is a useful starting point. Froglet shares a service: something another person’s AI assistant can query or run. You host it on a computer you control. It is not a hosted file-storage account.
Do I need to know how to code?
For a small data catalog, your assistant can help prepare it. Today you use a file-capable assistant such as Codex or Claude Code on Apple Silicon macOS or Linux. Installation and publishing still require your review.
What is private, and what becomes public?
You choose the data included in the service. Public services are accessible to others; invited services need a separate access credential. A service’s description, example and selected package information may still be public.
Can I charge for what I share?
Paid services are part of the vision, and payment integrations exist with different levels of maturity. The guided publishing journey starts free and is not yet a ready-to-use sales and billing service.
How do organisations run federated learning without sharing raw data?
Each organisation runs its own Froglet node over its local dataset and exposes a locked WebAssembly training service. A coordinator using Flower, NVIDIA FLARE or the froglet_fl SDK dispatches weights, receives only updated weights and signed receipts, and aggregates them with FedAvg, MultiKrum and Shapley attribution.
Can someone run their own program on my computer?
Yes, if you choose to offer bounded computation. A recipient can supply a compatible WebAssembly program and input; your node runs it with agreed memory, time and execution limits. It does not let arbitrary programs use your files or network.
Where does my service run?
On your computer or infrastructure. That machine needs to stay awake, online and running Froglet. Managed always-on hosting is not available yet, and this is a public beta.
Does Froglet enforce every part of an agreement?
It can apply supported technical controls such as selected data, invitation access and execution limits. Signed records make an exchange inspectable, but they do not prove a result is correct or automatically enforce legal terms, data-use promises or an uptime SLA.
Start small
What do you have that someone else could use?
One catalog. One lookup. One useful calculation.
Share something usefulExplore servicesPublic beta · Apache-2.0 · a project of Alithea Bio
Built in the open
Froglet is built by Armanas Povilionis and developed in the open, with documentation, a protocol specification and community discussions on the Froglet website and GitHub.
A project of Alithea Bio
Froglet is a project of Alithea Bio. Alithea Bio’s core business remains HLA immunopeptidomics for cancer immunotherapy, supported by bioinformatics and AI; Froglet is a separate open-source effort for sharing data and tools.