Proxies and gateways

Pipelock vs AWS AgentCore Gateway

A managed AWS gateway that turns APIs, Lambda functions, and services into MCP tools with ingress and egress auth, next to a self-hosted agent firewall that inspects payloads and signs receipts.

At a glance

Pipelock source AgentCore Gateway
Job Agent firewall. Mediates HTTP, WebSocket, and MCP traffic routed through it, scans it for secret leaks, prompt injection, SSRF, and tool poisoning, and can emit signed action receipts for mediated decisions when a signing key is configured. Managed AWS endpoint for agentic traffic: converts APIs, Lambda functions, and services into MCP-compatible tools, fronts other agents and HTTP services, routes model requests, and manages OAuth on the way in and the way out.
Enforcement point Network path, outside the agent process AWS-managed endpoint between agents and the tools, agents, and models behind it
Source Open source, Apache-2.0 core; Enterprise under ELv2 Closed source; AWS managed service
Pricing shape Free core; paid Pro and Enterprise tiers AWS usage pricing
Runs as Single Go binary, self-hosted; container and Helm Provisioned in an AWS account; no binary to run
Pick AgentCore Gateway

Your agents and tools live in AWS, you want APIs and Lambda functions exposed as MCP tools without writing servers, and OAuth on both sides handled for you.

Pick Pipelock

You need payload inspection on the agent's routed path, including outside AWS, and signed records where receipt signing is configured.

Run both

AgentCore Gateway fronts the AWS tools and agents. Pipelock can sit next to the agent and inspect traffic routed to it, including traffic sent to that gateway.

Want the runtime boundary, not just another checklist?

The short version

Pipelock is an open-source agent firewall. It inspects HTTP, WebSocket, and MCP traffic routed through it on the agent’s network path for credential leaks, prompt injection, SSRF, and tool poisoning. With a signing key configured, it can emit signed action receipts for mediated decisions. It can run beside the agent wherever the deployment routes traffic through it.

AgentCore Gateway is part of Amazon Bedrock AgentCore. Its developer guide describes one endpoint that routes and secures agentic traffic. It converts APIs, Lambda functions, and existing services into MCP-compatible tools from OpenAPI, Smithy, or Lambda definitions, and fronts other agents and HTTP services as passthrough targets, including A2A. It also routes model requests across providers and manages OAuth authorization for who may call it and how it calls upstream. A policy feature evaluates tool calls at the gateway.

AgentCore Gateway is managed plumbing and authorization inside AWS. Pipelock is inspection and evidence next to the agent.

Feature comparison

FeaturePipelockAgentCore Gateway
ArchitectureSelf-hosted single binary on the agent’s pathManaged AWS endpoint
What it frontsTraffic routed through it: HTTP, HTTPS CONNECT (payloads only with TLS interception on), WebSocket, MCP, A2AMCP tools built from APIs, Lambda, and services; other agents and HTTP services; model providers
Tool creationNo; it inspects tools that existYes, from OpenAPI, Smithy, and Lambda
AuthorizationTool policy via configOAuth ingress and egress; policy evaluation on tool calls
Model routingNoYes
Credential scanning (DLP)65 patterns, encoding-aware, on mediated trafficNot described in the pages read
Prompt injection detectionDeterministic patterns with multi-pass normalization on mediated trafficNot described in the pages read
Tool poisoning and rug-pull driftFingerprinting and drift detection on mediated MCP trafficNot described in the pages read
SSRF protectionMediated private-IP, metadata, and DNS rebinding checksNot described in the pages read
EvidenceSigned receipts when configured, verifiable offline with the operator’s keyAWS logging
Runs outside AWSYesNo
Source and pricingApache-2.0 core, freeClosed; AWS usage pricing

Where AgentCore Gateway is stronger

Tool creation from what you already have. An OpenAPI spec or a Lambda function becomes an MCP tool without writing a server.

Managed auth on both sides. OAuth for callers and for upstreams, run by AWS.

One endpoint for tools, agents, and models. Composition and model routing in the same place, with policy evaluation on the calls.

Nothing to operate. It is a service in your account.

Where Pipelock is stronger

Content inspection. Pipelock applies encoding-aware DLP to mediated bodies and arguments, injection normalization to mediated responses, tool-description fingerprinting and drift detection to mediated MCP traffic, and SSRF checks before DNS. The AgentCore pages read describe authorization and translation, not payload scanning.

Portability. Pipelock runs on a laptop, a CI runner, or any cloud, and inspects the agent’s whole path, not only calls to one gateway, when the deployment routes all egress through it (containment or a container with the proxy as its only route).

Evidence you hold. With a signing key and allow-path receipt enforcement configured, Pipelock refuses to forward an allowed mediated request whose signed receipt cannot be emitted. Its verifier checks signed receipts offline against your key.

Architecture difference

Agent -> AgentCore Gateway (OAuth in, translate, policy, OAuth out) -> APIs, Lambda, agents, models
Agent -> Pipelock (inspect, decide, sign)                           -> AgentCore Gateway and everything else

Put the gateway in front of the AWS estate and the firewall next to the agent.

Further reading

Sources checked

Third-party descriptions on this page come from the public materials below, read on the dates shown. Features and pricing change; check the current documentation before you decide.

Third-party product names and marks belong to their owners. PipeLab is not affiliated with, sponsored by, or endorsed by the makers of any product compared on this page. Descriptions of other products come from their own public materials on the dates listed above and reflect PipeLab's reading of them. If something here is wrong or out of date, tell us and it will be corrected.

Frequently asked questions

What's the difference between Pipelock and AgentCore Gateway?
AgentCore Gateway is a managed AWS endpoint. It converts APIs, Lambda functions, and existing services into MCP tools, fronts other agents and HTTP services, routes model requests, and handles OAuth authorization on the way in and out, with a policy feature that evaluates tool calls. Pipelock is a self-hosted agent firewall that inspects the content of HTTP, WebSocket, and MCP traffic routed through it. With a signing key configured, it can emit signed action receipts for mediated decisions. One is managed plumbing and authorization in AWS. The other is inspection and evidence next to the agent.
Does AgentCore Gateway inspect tool arguments or responses for injection and leaks?
Its developer guide describes translation, composition, authorization, model routing, and policy evaluation. Content scanning of tool arguments or responses, tool-description drift detection, and signed offline-verifiable receipts are not described in the pages read for this comparison.
Can I use both?
Yes. AgentCore Gateway fronts the tools and agents inside AWS; Pipelock can run next to the agent and inspect traffic the agent routes through Pipelock, including traffic sent to that gateway.

Want the runtime boundary, not just another checklist?

See all comparisons →