F5 AI Guardrails (formerly CalypsoAI)
CalypsoAI's runtime AI guardrails, sold by F5 since the September 2025 acquisition
Overview
F5 AI Guardrails is the runtime AI-security product F5 built from CalypsoAI, acquired in September 2025. It sits between applications and AI models or agents, blocking prompt injection and jailbreaks, catching sensitive-data leakage and logging every enforcement decision, in public cloud, private cloud or air-gapped on-prem. Red-team testing is sold separately as F5 AI Red Team.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 12 Oct 2026 and may be out of date or incomplete.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.
- No public price list
- The $100,000-a-year figure on AWS Marketplace is a stand-in for the private offer, not a price
- No self-serve trial listed
Use cases
What you can produce with F5 AI Guardrails (formerly CalypsoAI)
- Runtime blocking of prompt-injection and jailbreak attempts against models and agents
- Detection and redaction of sensitive data before it reaches a model
- Custom guardrails written in natural language
- Prebuilt controls for personal, payment and health data and EU AI Act categories
- Blocking of unauthorised agent tool calls and actions
- Audit-ready logs explaining each decision, exportable to a SIEM
- Deployment in AWS, Azure, Google Cloud, a private cloud or air-gapped on premises
F5 AI Guardrails is CalypsoAI's runtime layer under F5 ownership: policy enforcement on prompts, responses and agent tool calls, with prebuilt data-protection and EU AI Act controls, natural-language custom policies and explainable logs. It can run fully air-gapped, which matters to regulated buyers.
It is an enterprise purchase with no public price, transacted as a private offer, and the red-teaming CalypsoAI was known for is now a separate product. It suits large organisations, particularly F5 customers; small teams should start with open-source guardrail libraries.
What we found
F5 AI Guardrails is what CalypsoAI became. F5 announced the deal on 11 September 2025 at $180 million in purchase consideration and closed it on 26 September; its quarterly filing records $145.2 million paid in cash, with CalypsoAI becoming a wholly owned subsidiary. Earlier versions of this page dated the acquisition to 2026. The old calypsoai.com domain now redirects to F5's product page, and the product went generally available under the F5 name on 14 January 2026, alongside a separate product, F5 AI Red Team.
Guardrails is a runtime layer that sits between applications and AI models or agents. F5 says it blocks prompt injection and jailbreaks, detects sensitive-data leakage, moderates harmful output and can stop unauthorised agent tool calls. Policies can be written in natural language, and there are prebuilt controls for personal, payment and health data and EU AI Act categories. Every enforcement action is logged with its reasoning. It runs in AWS, Azure and Google Cloud, in a private cloud or fully air-gapped on premises, which is the realistic option for banks and public-sector buyers with data-residency rules.
The automated attack testing CalypsoAI was known for now lives in AI Red Team, which F5 sells separately; its findings can feed Guardrails policies. F5 says the attack library grows by more than 10,000 patterns a month, a vendor figure, as is the claim that both products are deployed at Fortune 500 firms in financial services and healthcare.
There is no price list. The AWS Marketplace listing says pricing is configured by F5 as a private offer, and its $100,000 twelve-month figure is described there as a placeholder. No self-serve trial for Guardrails appears on F5's trials page. Before the deal, CalypsoAI raised over $40 million from investors including Paladin Capital Group and Lockheed Martin Ventures.
This suits large regulated organisations, especially ones already running F5 infrastructure. Small teams will find open-source guardrail libraries a cheaper starting point.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
Frequently asked questions
About this listing
This entry was compiled from publicly available data including F5 AI Guardrails (formerly CalypsoAI)'s official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with F5 AI Guardrails (formerly CalypsoAI) unless explicitly stated.
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 F5 AI Guardrails (formerly CalypsoAI) directly →
Spotted something out of date? Suggest an update →
Alternatives to F5 AI Guardrails (formerly CalypsoAI)
More in Security & Safety
