Crovia — The Global Standard for AI Training Data Attribution. Cryptographic proofs, absence verification, and automatic royalty settlement.
Project description
Crovia Core Engine (Open Core)
If AI is trained on your data, there should be a receipt.
Crovia is an offline-verifiable evidence engine for declared AI training data.
It does not accuse.
It does not infer intent.
It does not enforce compliance.
Crovia produces deterministic artifacts that anyone can verify independently.
Install
pip install crovia
Requires Python 3.10+. No external dependencies for the core pipeline.
Read this first (30 seconds)
If you read only one thing, read this:
Crovia turns a declaration into a closed, verifiable evidence capsule.
Nothing more. Nothing less.
No trust required.
No network required.
No hidden logic.
What Crovia produces (CRC-1)
Crovia generates a CRC-1 Evidence Pack — a closed set of files that fully describe:
- what was declared
- what was produced
- how integrity can be verified
Each CRC-1 pack contains:
receipts.ndjson— declared training receiptsvalidate_report.md— deterministic validation resulthashchain.txt— integrity hash-chaintrust_bundle.json— normalized trust summaryMANIFEST.json— authoritative artifact contract
All files are offline-verifiable.
Try it (single command)
Generate CRC-1 Evidence Packs
# Full pipeline — receipts in, evidence pack out
crovia run --receipts examples/minimal_royalty_receipts.ndjson --period 2025-11 --budget 1000000 --out out_crc1
This creates a fully self-contained evidence capsule in out_crc1/.
Disclosure Scanner — check a model's disclosure gaps
# Scan a HuggingFace model for missing training data declarations
crovia oracle scan meta-llama/Llama-3-8B
crovia oracle scan mistralai/Mistral-7B-v0.1
Evidence Wedge — check a directory for evidence artifacts
crovia wedge scan # scan current directory
crovia wedge scan --path ./my-project
crovia wedge status # one-line status
crovia wedge explain # what artifacts Crovia looks for
Other commands
crovia check <receipts.ndjson> # validate receipts (real)
crovia refine <receipts.ndjson> # fix share_sum / rank issues
crovia pay <receipts.ndjson> --period YYYY-MM --budget N # compute payouts
crovia bundle --receipts X --payouts Y # assemble trust bundle
crovia sign <file> # HMAC-sign any artifact
crovia trace <file> # generate / verify hashchain
crovia explain <file> # inspect any Crovia JSON/NDJSON
crovia license status # check tier (OPEN / PRO)
crovia bridge preview <model> # PRO capability preview
crovia mode show # show CLI config
crovia legend # full command reference
crovia scan(attribution spider) requires the FAISS corpus index — not yet in open core.
Runcrovia scan <file>for details.
- No network
- No secrets
- Fully deterministic
Inspect the artifacts
# Linux / macOS
ls out_crc1
cat out_crc1/MANIFEST.json
# Windows
dir out_crc1
type out_crc1\MANIFEST.json
MANIFEST.json defines exactly which files must exist.
Nothing implicit.
Nothing hidden.
Verify evidence (offline, by anyone)
Verification requires only the files themselves.
Example:
crovia-verify out_crc1
Expected result:
[OK] All artifacts present
[OK] trust_bundle JSON valid
[OK] Hashchain verified
[OK] CRC-1 VERIFIED
If verification fails, the evidence is invalid.
No trust assumptions.
No authority required.
Design principles
- Offline-first
- Deterministic
- No attribution claims
- No enforcement logic
- Evidence > opinions
Crovia produces facts, not judgments.
Where to see real evidence
Crovia Open Core does not ship conclusions.
All public, inspectable evidence generated with this engine lives here:
https://github.com/croviatrust/crovia-evidence-lab
That repository contains:
- reproducible CRC-1 capsules
- offline-verifiable artifacts
- neutral semantic observations (DSSE)
- presence / absence observations (Spider)
If you want to see results, go there.
If you want to reproduce them, stay here.
Source
https://github.com/croviatrust/crovia-core-engine
License
Apache-2.0
CroviaTrust
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file crovia-1.2.1-py3-none-any.whl.
File metadata
- Download URL: crovia-1.2.1-py3-none-any.whl
- Upload date:
- Size: 36.9 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.11.4
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
c4a239791bec21a9d491e6f00a7bf4bfd7ca834a82186851909bb38715417fac
|
|
| MD5 |
a990db6086b4a64f2493084596290540
|
|
| BLAKE2b-256 |
6b8696595c8a18c8dc09559964a873ae674d7f6d5f8a1e2c4e04af679bbf57d8
|