Lasso Security

Discovery, red teaming and runtime guardrails for AI agents

Security & Safety Enterprise Has API
Researched · Published · Reviewed
RECATOOLS Score
7.2 / 10
Capability
8.5
Value for money
6
Ease of use
6.5
ASEAN readiness
5.5
API quality
7.5
Founded
2023
HQ
Tel Aviv, Israel
Users
A growing roster of enterprise and public-sector customers, including via Lasso Federal
Launched
Emerged from stealth Nov 2023 with $6M seed
Developer
Elad Schulman (CEO), Ophir Dror (CPO), Lior Ziv (CTO), Yuval Abadi (COO)

Overview

Lasso Security is a Tel Aviv-based platform for securing generative AI applications and agents, covering discovery and AI inventory, posture management, automated red teaming, runtime enforcement and detection. It raised $30 million in September 2026 and sells mainly by quote, with a public-sector arm, Lasso Federal LLC, in the US.

Advertisement

Pricing

Pricing shown for reference only. These figures reflect RECATOOLS research as of 14 Oct 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.

Secured Gateway for LLMs
$50,000/year
AWS Marketplace, 12-month contract
  • Annual commit
  • Additional usage at $0.01 per unit
  • Private offers via sales

Use cases

Detecting and blocking sensitive data exfiltration through employee use of ChatGPT, Copilot, or Claude Automated red-teaming of internal LLM-powered applications before production deployment Shadow AI discovery across an enterprise — identifying which AI tools are in use without IT approval Runtime protection for production AI agents running on Bedrock, Vertex AI, or Azure OpenAI Public-sector GenAI governance and compliance enforcement (FedRAMP High, ITAR, CJIS environments via Lasso Federal)

What you can produce with Lasso Security

  • An inventory of AI agents, models, system prompts and tools across the organisation
  • Automated red-team findings from multi-turn agentic attacks against your own applications
  • Inline policy enforcement at the proxy, API or AI gateway layer
  • Alerts on AI threats mapped to MITRE and OWASP, with attack context and impact
  • Posture and supply-chain gap analysis aligned to NIST and OWASP frameworks
  • Visibility into employee use of unsanctioned AI tools
Advertisement

ASEAN Perspective

Lasso Security in Southeast Asia

Singtel Innov8 — the Singapore telco's strategic venture arm — is a named investor, which signals some ASEAN institutional awareness of the platform and may facilitate regional introductions. However, no publicly documented ASEAN end-customers or regional offices have been confirmed as of mid-2026, and Lasso's marketing, compliance references, and case studies are overwhelmingly US- and Israel-centric. ASEAN enterprises with GenAI governance mandates (particularly under Singapore's MAS FEAT Principles or emerging AI governance frameworks in Indonesia and Thailand) would find Lasso technically capable but should budget for a heavier integration and onboarding effort without local presence or regional support SLAs.

RECATOOLS Verdict

Lasso covers the full lifecycle of an AI deployment, from finding unsanctioned agents to red teaming them and enforcing policy inline, and its new CPU-based LEAP guardrail is a credible answer to the cost of putting an LLM judge in front of every request. Its speed and accuracy figures are its own benchmarks, so test them in a proof of value.

Pricing is by quote, with one public anchor: $50,000 a year for its gateway on AWS Marketplace. Its commercial weight is in the US and Europe; ASEAN buyers will find no local references yet.

Desk research by RECATOOLS from public sources, written with AI assistance — not a hands-on test. How we review

What we found

Lasso Security is a Tel Aviv company building security for generative AI applications and agents, set up in 2023 by chief executive Elad Schulman and chief product officer Ophir Dror. It has not been acquired. In September 2026 it announced a $30 million round led by ClearSky Advisors, with Entrée Capital, which led its $6 million seed, adding to its stake and iAngels, Singtel Innov8, Mindset and Swish Data joining. SiliconANGLE put total funding above $37 million.

The platform runs as a loop: discovery and an AI bill of materials covering agents, models and system prompts; posture management mapped to NIST and OWASP; automated red teaming from what Lasso calls a library of more than 3,000 attacks; inline runtime enforcement at a proxy, API or AI gateway; and detection and response. The September launch added LEAP, a transformer-free guardrail that runs on ordinary CPUs and, by Lasso's own benchmarks, returns a verdict in under five milliseconds, plus RAPID, a self-hosted model for the judgement calls. Lasso also claims 98.6% detection accuracy; none of these figures has been tested outside the company.

Customers named in the round coverage include BMW, eToro, Fiverr and the US Department of Homeland Security, and Lasso said revenue grew more than 500% in a year without giving an absolute number. In July 2025 it set up Lasso Federal LLC with reseller Swish Data, offering the product through AWS GovCloud to agencies with FedRAMP High and DoD SRG requirements; that describes a hosting environment, not a Lasso authorisation.

Pricing is quote-based. The one public figure is an AWS Marketplace listing for its Secured Gateway for LLMs at $50,000 for a 12-month commitment, plus $0.01 per unit of extra usage. Its open-source MCP Gateway is MIT-licensed but has had no commits since January 2026. Apart from the Singtel Innov8 stake, no Southeast Asian customer is documented.

Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.

Notable facts

  • Former Israeli Prime Minister Naftali Bennett sits on Lasso's board — he previously co-founded cybersecurity firm Cyota, which was acquired by RSA Security.
  • Lasso's open-source MCP Security Gateway, released in April 2025, reached 376 GitHub stars within months and was among the first tools to secure Model Context Protocol traffic for agentic AI.
  • The company positions itself as 'the Wiz of GenAI security' — a deliberate nod to Wiz's cloud-security playbook of comprehensive visibility and rapid enterprise adoption.
  • Lasso claims its runtime enforcement engine is 570x more cost-effective than cloud-native guardrails such as AWS Bedrock Guardrails or Azure AI Content Safety, based on its own benchmarks.

Frequently asked questions

Discover inventories every AI agent, model, and shadow AI tool across cloud and CI/CD pipelines. Assess analyses security posture, misconfigurations, and compliance gaps against NIST and OWASP. Test runs automated red-teaming using 3,000+ attack scenarios including multi-turn agentic attacks. Enforce deploys inline policy controls at the proxy, API, or AI Gateway layer in under 50ms. Protect provides continuous detection and termination of live AI threats mapped to MITRE and OWASP taxonomies.
The core commercial platform is proprietary and enterprise-licensed. However, Lasso maintains open-source MIT-licensed tooling on GitHub, most notably the mcp-gateway (a plugin-based security gateway for Model Context Protocol) and claude-hooks (prompt-injection defences for Claude Code). These serve as community on-ramps and do not replace the full commercial product.
Lasso developed what it calls the Intent Security Framework — a behavioural baseline approach that monitors not just what an AI agent does but what it intends to do, detecting anomalies before a malicious action completes. This goes beyond simple keyword or content filtering and is designed for multi-step agentic workflows where conventional guardrails produce too many false negatives.

About this listing

Researched on
Published on
Last reviewed

This entry was compiled from publicly available data including Lasso Security's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Lasso Security unless explicitly stated.

Data accuracy

Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.

For the latest details, please refer to Lasso Security directly →

Spotted something out of date? Suggest an update →

Advertisement