Regulatory intelligence as machine-readable agent objects
Five open source agents for legal, compliance, regulatory, case law and risk questions. Every answer carries its source and proof. No key, no contract, no pilot. You need no RAG, GPU or own rule interpretation.
Source code on GitHubMachine starting pointAgents
| Agent | Question | Coverage |
|---|---|---|
legalAgent | What applies? | Provision node with case law and evidence chain (SE, NO, US, EU) |
complianceAgent | What must we do? | Obligations from a company profile: country, industry, employees, turnover |
regulatoryAgent | What changed this week? | GDPR, DORA, AI Act, NIS2, CSRD, eIDAS, MiCA, DSA, Data Act, AMLR with Official Journal dates and the current consolidated EUR-Lex version |
caseLawAgent | How have courts decided? | Hash-chained case law lines for the EU, the US (incl. Supreme Court), Sweden and Norway |
riskAgent | What do we risk? | Risk objects, sanctions and supervisory decisions in one answer |
Trust by design
- Source on every answer. Source URL, server proof and a local
receipt_sha256of the exact response, usable as an audit trail. - No guessing. Missing data is returned as
not_in_register. - Stable identifiers. Object IDs and published hash formulas are append-only, so stored evidence stays valid.
- Daily verification. EUR-Lex, national courts and registers are checked every day.
- Zero dependencies. Python standard library and plain
fetchin TypeScript.
Quick start
git clone https://github.com/stigprivate-design/nova-open-agents cd nova-open-agents node --experimental-strip-types examples/risk-and-case-law.ts
import { regulatoryAgent, riskAgent, caseLawAgent } from "./index.ts";
const mica = await regulatoryAgent("mica"); // official dates + consolidated version
const risk = await riskAgent("gdpr"); // risk, sanctions, decisions
const cases = await caseLawAgent("sherman-act", "us");
console.log(mica.proof, risk.objects, cases.objects);
from nova import regulatory_agent, risk_agent
print(regulatory_agent("dora")["objects"]["official_text"])
print(risk_agent("gdpr")["proof"])
Open endpoints
Base: https://legal.exploreworldai.com/api/public/v1
| Endpoint | Used by |
|---|---|
/provision-node?id= | legalAgent |
/decision?country=&industry=&employees= | complianceAgent |
/regulation-week?reg= | regulatoryAgent |
/praxis-line?act=&jurisdiction= | caseLawAgent |
/risk-assess?act=, /sanctions?act=, /enforcement? | riskAgent |
/duty-objects, /graph/obligation|provenance|action | SDK |
MCP: https://legal.exploreworldai.com/.well-known/mcp.json
Fair use
Open access is free. Identify your integration with a clear User-Agent; the SDK sends NovaOpenAgent/<version>. Unknown high-volume crawlers are limited to 100 requests per minute.
Enterprise
The open agents and endpoints are free and stay free. Organisations that need contractual guarantees can agree on NovaCopilot Enterprise: service level agreement, dedicated support, higher volumes, private deployment and onboarding for compliance, risk and audit teams. Contact: stig@valkiv.com