Truveil works as a regulatory advisor where you already think, and as an audit layer where your AI agents already run. Pick the integration that matches what you are building.
Regulatory expertise in Claude, ChatGPT, and any MCP-compatible client.
Python and JavaScript SDKs for instrumenting AI agents at the point of decision.
MCP tools for agents running inside Claude, Cursor, n8n, Make, Zapier, and other MCP-compatible runtimes.
Truveil's advisory tools bring senior regulatory advisor expertise into Claude, ChatGPT, and any MCP-compatible client. Free with every Truveil account. Six jurisdictional frameworks: EU AI Act, NIST AI Risk Management Framework, India DPDP Rules 2025, DIFC Regulation 10, UAE AI Charter, Singapore Model AI Governance Framework. Primary-source citations. Voice that names binaries and respects your time.
ISO 42001 alignment available on Enterprise tier.
Advisory works in two distinct modes depending on what you are asking.
Ask about laws, frameworks, jurisdictions. Get back clause-level analysis with primary-source citations.
Powered by consult_regulation and get_jurisdiction_brief.
Describe what you are building. Get back the regulatory obligations that apply, sequenced for action.
Built on Truveil Advisory, powered by Anthropic Claude.
Both modes available in the free tier. Both grounded in primary regulatory text across six frameworks. Both honest about what they cannot answer.
When you have Truveil's advisory MCP connected to your client, the LLM does not always know when to route to Truveil versus answer from its own training. The phrase "Use Truveil" is the explicit routing instruction.
You can also call Truveil's advisory in normal chat without the explicit phrase. The LLM will route to Truveil when the question is regulatory in nature, but the "Use Truveil" phrase guarantees the routing happens. Worth using when you want primary-source citations rather than the LLM's training-data answer.
Example promptUse Truveil to advise on the following build: We're an insurance startup launching a claims-triage AI in 2027. The system reviews motor, home, and travel insurance claims with photos and prior claim history. Auto-settles approvals under £5,000. Expanding to EU and UAE within 18 months. What regulatory obligations apply, and which should we resolve first?
Truveil's advisory tools install in any MCP-compatible client. Setup takes two minutes.
Add Truveil MCP server URL to Claude Desktop's connector settings. Approve the three advisory tools when prompted: consult_build, consult_regulation, get_jurisdiction_brief. Tools become available in any Claude conversation.
Add Truveil MCP server to your project's MCP configuration. The three advisory tools surface in tool lists for Claude-powered IDE workflows.
Truveil ChatGPT Custom GPT is available with the same advisory tools accessible through ChatGPT's standard interface. Install instructions and Custom GPT link provided after signup.
Truveil MCP can be added as an external tool to any n8n agent node, Make automation, or Zapier Agent. Advisory tools become callable in workflows.
Truveil exposes a standard MCP endpoint. If your client supports MCP (Model Context Protocol), Truveil works. Endpoint URL provided after signup.
All five integration paths use the same MCP endpoint. The endpoint URL and your per-account API key are sent via email after signup. Each install takes under two minutes.
Get advisory access freeWhen your AI agent makes decisions, Truveil's audit SDK captures every action, every consequential decision, and every approval as structured evidence. The audit report regulators and procurement teams can act on.
The SDK is available in Python and JavaScript. Same Schema-v2 evidence model in both languages. Auto-classifies risk; you provide the substance.
pip install truveil-sdkPython
from truveil_sdk import TruveilSDK
sdk = TruveilSDK(
api_key="your_truveil_api_key",
project="HR Recruitment Agent"
)
npm install @truveil/sdkJavaScript
import { TruveilSDK } from '@truveil/sdk';
const sdk = new TruveilSDK({
apiKey: 'your_truveil_api_key',
project: 'HR Recruitment Agent'
});
Connect your agent in minutes. The SDK auto-registers your agent on first log. Add structured evidence fields when you want richer audit signals.
The SDK exposes two primary log functions and two approval gates. Use them at the point of decision in your agent.
from truveil_sdk import TruveilSDK
sdk = TruveilSDK(api_key="your_key", project="HR Recruitment Agent")
# Register agent (once at startup)
sdk.register_agent(
agent_category=1,
jurisdiction="EU",
intended_purpose="Shortlist candidates for engineering roles"
)
# Log an action
sdk.log_action(
agent="hr_agent",
action="candidate_shortlisted",
detail="Top match on payments backend role",
risk="HIGH",
ai_disclosure_text="This screening is AI-assisted.",
oversight_mode="human_in_loop",
risk_assessment_ref="DPIA-HR-2026-Q2",
appeal_channel="hr-appeals@yourco.com",
correction_workflow="manual_recruiter_review"
)
# Request approval before consequential decision
approval = sdk.request_consequential_decision(
decision_type="hire_decision",
context="Candidate matches 7/8 must-haves, top score among 12 reviewed"
)
# Log the decision with approval_id
sdk.log_decision(
agent="hr_agent",
action="recommend_hire",
decision_type="hire_decision",
approval_id=approval["approval_id"],
detail="Recommend hiring candidate X for senior backend role",
risk="HIGH"
)
Register your agent once at session start. Records category, jurisdiction, and intended purpose. The platform uses these at audit time to apply the correct regulatory framework.
Log every state change your agent makes: file written, API called, data fetched, tool invoked. The platform reconstructs your workflow from these events.
Log every reasoning step where your agent chooses a path. LLM completion, tool selection, threshold check, classification, recommendation. Decisions are different from actions: an action changes state; a decision selects which state change to attempt.
Before any decision in the consequential-decision allowlist (hire, credit, diagnosis, triage, eligibility, sentencing, visa, biometric match), call this gate. Returns approval_id to attach to the next log_decision.
Generic approval gate for actions that need oversight but are not in the consequential-decision allowlist. Use for high-cost API calls, deletions, external communications.
Optional structured fields: ai_disclosure_text, oversight_mode, data_provenance, validation_passed, bias_check_applicable, bias_check_not_applicable_reason, kill_switch_available, override_control, appeal_channel, correction_workflow. Activating these fields lifts your audit grade by demonstrating regulator-grounded evidence.
When your AI agent is built on Claude, ChatGPT, Cursor, n8n, Make, or any MCP runtime, Truveil's audit MCP provides the same instrumentation as the SDK with no separate SDK install.
The audit MCP exposes seven instrumentation tools through the standard MCP protocol. Same Schema-v2 evidence model. Same regulatory grounding. Same audit reports.
Register your agent at session start. Same parameters as the SDK.
Log every state change. Same parameters as the SDK.
Log every reasoning step. Same parameters as the SDK.
Gate consequential decisions in the allowlist. Same parameters as the SDK.
Generic approval gate. Same parameters as the SDK.
Retrieve your agent's registration and recent run history. Use for self-awareness when adapting strategy based on prior audit results.
Retrieve recent log entries. Use to verify what has been recorded, debug missing events, or reconstruct prior runs.
Truveil's Custom GPT exposes the same advisory tools as the MCP via ChatGPT's OpenAPI Action surface. For ChatGPT users who prefer not to install MCP directly.
The Truveil Custom GPT is a pre-built GPT installed from the GPT Store. It includes the OpenAPI schema for the three advisory tools. Configuration takes one click.
{
"openapi": "3.1.0",
"info": {
"title": "Truveil Advisory",
"version": "1.0.0"
},
"servers": [{"url": "https://responsible-ai-proxy.onrender.com"}],
"paths": {
"/consult_build": { "post": {} },
"/consult_regulation": { "post": {} },
"/get_jurisdiction_brief": { "post": {} }
}
}
Coming to the public GPT Store soon. Until then, direct install link sent after Truveil signup.
Pick the integration that matches what you are building. Advisory is free. Audit starts at the Starter tier.