A

Authensor

Authensor is an open-source security governance layer for AI agents, adding policy gates, content scanning, approvals and full audit trails—so teams can block, review or allow every tool call before it runs.
Rating:
5
Visit Website
AuthensorAI agent security guardrailsLLM tool-call policy engineprompt-injection detectionself-hosted AI governanceLangChain security middlewareagent kill-switchAI approval workflow

Features of Authensor

Policy gate & decision control before any external action
ALLOW / BLOCK / ASK outcomes plug straight into approval flows
Content scan flags prompt injection and PII leakage
Tamper-proof audit receipts with append-only hash chain
Real-time budget caps, rate limits & anomaly alerts
Kill-switch + canary/shadow mode for safe policy rollouts
Multi-party approvals, webhooks, TTL & delegation chains
MIT open-source, self-host or use managed dashboard
Works with LangChain, Claude, OpenAI, CrewAI, MCP, Vercel AI SDK

Use Cases of Authensor

Add a unified security gate before production AI agents hit APIs or tools
Force human approval when an agent touches DBs, files or payment endpoints
Block prompt-injection and data-exfiltration in support or ops bots
Give security teams tamper-proof logs for post-incident forensics
Centrally manage permissions across multi-team agent fleets
Embed policy checks in CI/CD to catch risky agent updates early
Keep data on-prem with self-hosted deployment while still governing agents
Act as a neutral governance layer in multi-agent orchestrators

FAQ about Authensor

QWhat is Authensor?

Authensor is an open-source guardrail and governance layer that sits in front of AI agents, enforcing policies, scanning inputs/outputs, and recording every decision for audit.

QWhich problems does it solve?

It stops agents from calling tools they shouldn’t, blocks prompt-injection attacks, prevents sensitive data leaks, and gives you a full audit trail.

QWhich agent frameworks are supported?

LangChain, OpenAI Assistants, Claude, CrewAI, Vercel AI SDK and MCP; SDKs make integration one-line.

QCan I self-host Authensor?

Yes—the core is MIT-licensed and ships as Docker containers. A managed dashboard is optional.

QHow does the approval workflow work?

High-risk actions return an ASK verdict; you route it to humans via Slack, email or ticket, with webhooks, timeouts and multi-party sign-off.

QDoes it provide audit logs?

Every decision is written to an append-only, hash-chained ledger you can stream to SIEM or store offline.

QIs Authensor free?

The open-source engine is free; the cloud dashboard and enterprise support are paid—see the pricing page for details.

QIs Authensor SOC 2 or ISO certified?

The docs say the architecture maps to SOC 2 controls, but no third-party certification is claimed—verify current status with the vendor.

Similar Tools

A

AgentAnchorAI

AgentAnchorAI is an enterprise-grade governance platform for AI agents, delivering pre-execution gatekeeping, dynamic authorization and full audit trails—so teams can run multi-agent workflows with built-in risk controls.

F

FencioAI

FencioAI delivers runtime security and governance for AI agents—helping teams benchmark before launch, enforce policies in production, and maintain a full audit trail to manage risk with confidence.

i

iAgentic AI

iAgentic AI is an enterprise-grade AI control plane for decision governance—unifying policy enforcement, approval workflows and audit trails across multi-model, multi-system environments.

E

ExecLayer AI

ExecLayer AI delivers an enterprise-grade execution-governance layer for AI Agents. It enforces approval workflows, policy controls and full audit trails, letting teams deploy AI in live processes with confidence.

A

AgentSec

AgentSec delivers zero-trust security for AI agents—covering access control, audit trails, and data protection—to cut privilege-escalation and data-exfiltration risks from development through runtime.

R

Runshift

Runshift is an Agent Control Plane for production. It lets teams operate multi-AI-agent systems with confidence by gating critical actions, enforcing human approval, routing requests and keeping a full audit trail.

A

Auth0 AI

Auth0 AI gives AI agents enterprise-grade authentication and authorization: delegate API access on behalf of users, vault tokens, and enforce fine-grained permissions so teams can ship controllable, audit-ready agent workflows.

L

LangGuard AI

LangGuard AI is a unified AI control plane for enterprise IT and security teams to discover, approve, monitor and audit every AI asset—agents, models, tools and data—through one governance layer.

A

Agent Action Firewall

Agent Action Firewall is an enterprise-grade AI-agent security platform that delivers policy-based governance, human-in-the-loop approval, and tamper-proof audit trails—covering the full discover-govern-comply lifecycle.

A

Alterion AI

A runtime control plane for AI agents—observe in real time, govern via a semantic layer, and give your team full visibility, audit trails, and policy enforcement across multi-cloud and distributed environments.