Arcjet vs Lasso Security

How do Arcjet and Lasso Security differ?

Lasso Security is an enterprise AI security platform that covers discovery, posture, red teaming, and runtime protection, with enforcement at a proxy, API, or AI gateway layer and through coding-agent hooks. Arcjet is an AI agent runtime security platform that enforces policy in the path of each action: in your code for custom agents, and in the hooks of Claude Code, GitHub Copilot, Cursor, and OpenAI Codex, with no proxy to route traffic through. The key difference is the enforcement point for your own applications. Lasso routes model traffic through a proxy or gateway layer, and Arcjet decides at the tool call in your code. Choose Arcjet if you want one authored policy across four coding agents and your own agents, decided before the action runs, on self-serve pricing. Choose Lasso if your priority is employee generative AI use in the browser, red teaming, and posture management from one enterprise vendor.

Arcjet publishes this comparison. Competitor details come from Lasso Security's public documentation, reviewed on September 25, 2026. Products change, so check the linked sources before you decide.

Arcjet and Lasso Security at a glance

Arcjet enforces in the agent's hook and in your code, and Lasso Security enforces at a proxy, API, or AI gateway layer and through coding-agent hooks. The following table compares the two products across 11 areas.

What is Lasso Security?

Lasso Security is an enterprise AI security platform that its home page describes in five parts: discovery with an AI bill of materials (AI-BOM), AI security posture management, automated red teaming, runtime enforcement, and AI detection and response. The platform page describes a red-teaming library of more than 3,000 attacks mapped to the OWASP Top 10, and runtime enforcement "inline at the proxy, API, or AI Gateway layer."

For coding assistants, Lasso connects to agent lifecycle hooks and names four agents: Claude Code, Cursor, Codex, and OpenCode. Lasso's Claude Code article describes deployment through Claude for Enterprise managed settings, inspection of tool calls before they run, and scanning of content the agent ingests, including tool results and MCP server responses.

Lasso also discovers MCP servers across an organization and runs a gateway that filters MCP requests and responses. Lasso's open-source MCP Gateway, released under the MIT license, ships secret-masking, Presidio PII, and Lasso guardrail plugins. For employees who use public chatbots, Lasso offers a browser extension.

On September 2, 2026, Lasso announced LEAP, a guardrail that runs on standard CPUs with a stated decision latency under 5 milliseconds, paired with an LLM-as-judge called RAPID for complex policy decisions. The same announcement describes deployment in customer environments, including air-gapped networks, a $30 million Series A led by ClearSky, founding in 2023, and SOC 2 Type II certification.

Lasso's website doesn't publish pricing. An AWS Marketplace listing for Lasso's Secured Gateway for LLMs shows a 12-month contract at $50,000.

What is Arcjet?

Arcjet is an AI agent runtime security platform. Arcjet discovers the agents running in your organization, enforces policy on every action, prompt, and tool call, and keeps the evidence of what happened. Arcjet decides in the path of the action rather than on the network.

For coding agents, Arcjet enforces policy on Claude Code, GitHub Copilot, Cursor, and OpenAI Codex through the hooks each agent already fires, with no SDK and no code change. One policy applies across all four agents, and you choose when it runs: Tool call, Prompt, or Model switch. You install the hooks through managed settings or MDM, so developers can't remove them without administrator access.

Arcjet's 12 starter policies cover destructive commands, history rewrites, protected paths, credential access, piped installers, MCP and egress allowlists, destination threats, model allowlists, sensitive information, prompt injection, and npm publish. Evaluation runs at the edge in over 300 data centers, and a hook typically adds a few tens of milliseconds.

For custom agents, the JavaScript, TypeScript, Python, and Go SDKs call guard() on tool calls, queue consumers, MCP handlers, and other actions that have no HTTP request, and protect() on HTTP routes. Arcjet integrates with 15 agent frameworks, including the Claude Agent SDK, LangChain, LangGraph, Mastra, OpenAI Agents, the Vercel AI SDK, and Microsoft Agent Framework for Go. You write Guard policies in Rego or a builder, dry run them before they go live, and publish them from the Arcjet Console or MCP server, and changes take effect in real time.

Arcjet detectors include prompt injection, sensitive information with an SDK-local option that keeps the raw body in your process, destination threat analysis, token budgets, and rate limits. The same SDK covers bot detection, Shield WAF, and email validation for your web application. The Arcjet Console records every decision and session, and SIEM export to Datadog, Splunk, SentinelOne, Panther, and Amazon S3 is available on the Enterprise plan.

Key differences between Arcjet and Lasso Security

Arcjet differs from Lasso Security in four ways: where enforcement happens, coding-agent coverage and policy authoring, the handling of tool results, and coverage of data handling, web security, and pricing. The following sections cover each one.

Where enforcement happens

Lasso's runtime enforcement for applications sits at a proxy, API, or AI gateway layer, so traffic between your application and the model goes through that layer. Arcjet has no proxy. For your own agents, the Arcjet check is a function call at the tool, next to the code that performs the side effect, so the policy can read the tool name, the arguments, the user, and the account.

A tool call that never produces model traffic, such as a queue consumer acting on a stored instruction, still gets an Arcjet decision. For more information about this design, see AI agent security architecture.

Coding-agent coverage and policy authoring

Both products use the hooks that coding agents fire. Lasso's coding-assistant page names Claude Code, Cursor, Codex, and OpenCode, and Lasso's public documentation doesn't describe GitHub Copilot hooks. Arcjet applies one policy across Claude Code, GitHub Copilot, Cursor, and Codex.

You write the Arcjet policy yourself in Rego or the builder, start from one of 12 starter policies, and dry run it before it denies anything. A denial names the rule ID, the policy, and the decision, so a blocked developer and a reviewer can trace it. For more information, see how to secure AI coding agents.

Tool results

Lasso screens tool results before they reach the model, and Arcjet doesn't on coding agents. Lasso's Claude Code article describes scanning content that the agent ingests, including tool results and MCP server responses. Arcjet screens prompts and tool calls and records tool results as activity. If screening tool output is your main requirement, weigh this difference.

Data handling, web security, and pricing

Arcjet's SDK-local sensitive-information detection keeps the raw body in your process. Prompt injection detection runs in the Arcjet Cloud API, so the local-processing claim applies to sensitive-information inspection, not to every rule. For more information, see keeping security inspection local.

Arcjet also covers bots, rate limiting, and Shield WAF for the web application that hosts your agent, which Lasso's public documentation doesn't describe. Arcjet publishes its prices, from $25 per month, and offers a 15-day trial. Lasso's website doesn't publish pricing.

When to choose Lasso Security

Choose Lasso Security when you want browser, red teaming, posture, and gateway enforcement from one enterprise vendor. Lasso fits when any of the following apply:

• You need to govern employee use of public chatbots through a browser extension.
• You want automated red teaming and AI security posture management from the same vendor as runtime protection.
• You need guardrails that run inside your own environment, including air-gapped networks.
• Screening tool results and MCP responses before they reach the model is a firm requirement.
• You already route model traffic through an AI gateway and want enforcement there.

When to choose Arcjet

Choose Arcjet when you want one policy that decides each action before it runs, on coding agents and in your own agent code. Arcjet fits when any of the following apply:

• You want one policy that decides before the tool call runs on Claude Code, GitHub Copilot, Cursor, and Codex.
• You build your own agents and want enforcement at the tool call, with application context, and no proxy.
• You want to write and review policy as code in Rego, with dry run before live.
• You want bots, rate limiting, and WAF for the same application.
• You want published, usage-based pricing and a self-serve start.

Can you use Lasso Security and Arcjet together?

Yes, Lasso Security and Arcjet can run side by side. A common split is Lasso for browser-based employee use of generative AI, red teaming, and posture, and Arcjet for in-path enforcement on coding agents and in your own agent code. If both products install coding-agent hooks, decide which product owns each hook event, and test that a denial from either one stops the action.

What are the alternatives to Lasso Security?

Lasso Security alternatives include other enterprise AI security platforms, platform-native controls, and runtime enforcement in code. The following products cover overlapping ground:

• Arcjet provides in-path enforcement for custom agents in code and for Claude Code, GitHub Copilot, Cursor, and Codex through their hooks, plus web application security in the same SDK.
• Noma Security provides AI security posture management, access control, detection and response, and red teaming, with enforcement through agent hooks, gateways, SDKs, and APIs. For more information, see Arcjet vs Noma Security.
• Microsoft Agent 365 is a control plane that extends Defender, Entra, and Purview to agents, suited to Microsoft 365 organizations. For more information, see Arcjet vs Microsoft Agent 365.
• Check Point AI Agent Security provides agent discovery, configuration risk assessment, and runtime protection, documented as an early access release and built on Lakera. For more information, see Lakera alternatives.

For a wider map of the category, see AI agent security platforms.