Production API · SDK on PyPI

Action Control
for autonomous agents

Check posture. Gate risky actions. Prove execution.

Fits your existing agent stack.

Run the Loop → Get Started GitHub
Featured Glama MCP · 12 full / 8 hosted
∕∕ ACTION CONTROL
Request Posture Gate Execute Receipt
>>> action.requested

{
  "action_id": "act_24e9",
  "agent": "did:key:z6Mk...",
  "scope": "prod/api deploy",  REQUESTED
  "tool_access": ["github.write", "ci.run", "env.prod"],
  "receipt": "required"
}
Live
Production API
Latest
IPFS Anchor
W3C VC
Offline-Verifiable
Tamper-Evident
Audit Chain

>Start with one workflow. Expand when it matters.

Add posture checks, action gates, and signed receipts to the agent stack you already use.

For builders

Add action control to agent workflows

For developers and product teams that want posture checks, action gates, and signed receipts without replacing their existing stack.

  • Python SDK, REST API, and MCP server
  • CrewAI, LangGraph, AutoGen, OpenAI, Claude MCP, Paperclip
  • Local-first posture checks
  • Signed receipts for agent actions
Get started →
For operators

Gate risky agent actions in production

For teams that need approvals, monitoring, and verifiable evidence for high-risk agent actions.

  • Start with one critical action
  • Allow / require approval / block
  • Monitor trust drift and score drops
  • Export receipts clients can verify
See packages →

>Who AgentVeil is for

Built for teams shipping agents into real workflows — especially when agents can touch tools, data, code, or production systems.

AI agencies and studios

Give clients independent proof of what your agents did — without relying on dashboards, screenshots, or internal logs.

Multi-agent teams

Check agent posture before delegation, gate risky actions, and monitor drift over time. Add coverage quickly with @avp_tracked when you’re ready.

SaaS products with agent features

Add posture checks, action gates, and signed receipts around sensitive agent actions — without replacing your existing stack.

EU AI ACT

Client-facing or high-stakes workflows

Produce client-ready evidence for agent actions that can be verified independently. See Art. 9 / 12 / 13 / 14 / 50 mapping →

Technical readiness mapping only; AVP does not provide legal advice or compliance certification.

>How It Works

AgentVeil works in three stages: check posture before runtime, gate risky actions during execution, and verify receipts afterward.

Before
Posture Check
Check what the agent could do
Find risky tools and bypass paths
Recommend controls before runtime
During
Runtime Gate
Evaluate each risky action
Allow, require approval, or block
Monitor drift and trigger alerts
After
Signed Receipts
Record what happened
Export verifiable proof packets
Support audit and dispute review
Scope · what AVP is not

AgentVeil is not a policy engine, MCP gateway, or sandbox runtime. It adds posture checks, runtime gates, and signed receipts around your existing identity and governance stack.

>Independent proof for agent actions

AgentVeil creates signed receipts for risky agent actions — so clients, reviewers, and partners can verify what happened independently.

Step 1
Action requested
Step 2
Gate decision
Step 3
Execution recorded
Step 4
Receipt signed
Step 5
Offline verification
curl agentveil.dev/v1/reputation/{did}/credential?format=w3c

eddsa-jcs-2022 · RFC 8785 JCS · offline-verifiable with standard VC tooling

>Why not just logging?

Logs show what happened.
AgentVeil helps decide whether the action should happen — and produces signed receipts others can verify.

Logging alone
  • Records events after the fact
  • Controlled by the operator
  • No posture check before runtime
  • No action gate before execution
  • No offline-verifiable receipt
AgentVeil with logging
  • Checks agent posture before runtime
  • Allows, blocks, or requires approval for risky actions
  • Signs tamper-evident execution receipts
  • Exports proof clients and reviewers can verify
  • Supports audit, review, and dispute resolution

>Fits the stack you already have

Add posture checks, runtime gates, and signed receipts without replacing your framework, identity provider, or observability tooling.

Frameworks

CrewAI
LangGraph
AutoGen
OpenAI
Claude
📎 Paperclip
🐍 Any Python

Interfaces

Python SDK

pip install agentveil — Python SDK.

REST API

Any language. Full documentation & guides.

MCP Server

pip install 'agentveil[mcp]' — Claude Desktop, Cursor, Windsurf, VS Code.

Enterprise fit

Works with your existing identity stack

AgentVeil adds action controls and verifiable evidence around your current systems.

✔ MERGED

Composes with Microsoft AGT

AVPProvider merged into Microsoft Agent Governance Toolkit (PR #1010).

>Production API

Use production endpoints for reputation signals, advisory checks, runtime gate decisions, signed receipts, and audit verification.

Start with advisory checks. Add signed runtime gates for controlled actions when you’re ready.

Method Endpoint Access Description
GET/v1/reputation/{did}PublicReputation signal
GET/v1/reputation/{did}/trust-checkPublicAdvisory action check
GET/v1/reputation/{did}/credentialPublicSigned reputation credential
GET/v1/cardsPublicAgent discovery
GET/v1/audit/verifyPublicAudit chain verification
GET/v1/anchors/latestPublicLatest public IPFS anchor
POST/v1/attestationsSignedSubmit peer rating
POST/v1/attestations/batchSignedBatch ratings
POST/v1/agents/registerSignedRegister agent
POST/v1/runtime/evaluateSignedRuntime Gate decision
POST/v1/executeSignedControlled action execution
GET/v1/execution/receipts/{receipt_id}SignedOwner-only signed receipt lookup

SDK: github.com/agentveil-protocol/avp-sdk · PyPI

>Start in minutes

Try AgentVeil locally with a mock agent, then connect the same action workflow to production when you’re ready.

# Install the Python SDK
pip install agentveil

# Try instantly — no server needed
from agentveil import AVPAgent

agent = AVPAgent.create(mock=True, name="my_agent")
agent.register(display_name="My Agent")
rep = agent.get_reputation(agent.did)
print(rep)  # {'score': 0.75, 'confidence': 0.5, ...}
Connect to production server
# One line to auto-register and auto-attest an agent
# inside an existing workflow
from agentveil import avp_tracked

@avp_tracked("https://agentveil.dev", name="my_agent", to_did="did:key:z6Mk...")
def review_code(pr_url: str) -> str:
    return analyze(pr_url)

>Adopt AgentVeil in stages

Start free in development. Add a guided pilot around one critical action. Expand to multi-team or high-stakes environments when you need rollout support.

FREE

Build

For developers and teams adding local posture checks and signed proof artifacts in development

  • Python SDK + optional MCP tools
  • REST API + 7 framework integrations
  • W3C VC credentials, offline-verifiable
  • Local advisory decisions with can_trust()
  • Community support
pip install agentveil
RECOMMENDED

Pilot

Gate one critical action and prove execution before wider rollout

  • Everything in Build
  • Runtime gate on a critical action
  • Alerts + webhook integration
  • Audit trail export + dispute review
  • 30-day scope with rollout guidance
Discuss a pilot →
ENTERPRISE

Deploy

Scale action gates and independently verifiable receipts across production systems

  • Everything in Pilot
  • Multiple workflows, custom thresholds
  • Private deployment options
  • Compliance & security review path
  • Priority support + SLA
Discuss deployment →