AIRTA Red Team

AIRTA Red Team is a self-hosted tool for authorized runtime security assessments of LLMs and LLM-driven applications. It best suits red teams, pentesters, and AppSec or MLsec teams that want structured attack runs and evidence. Its unmaintained status makes it a poor choice for teams that need an actively supported tool.

Overview

AIRTA Red Team turns security playbooks into assessment suites, discovers target components and authentication, runs attacks, evaluates outcomes, and exports reports. Its runtime focus is useful for examining what an application exposes during interaction, but it does not perform source-code static analysis. The product calls for explicit permission before testing any system.

Deployment options include local Docker, environments behind a firewall, CI, and authorized black-box assessments of external endpoints. You can test through a Playwright browser runner or direct HTTP API transports, including file uploads and tool calls. This range supports both browser-based applications and API targets, though users must bring their own LLM API key for test generation and assessment.

Key features

Playbooks and attack strategies

Coverage includes OWASP LLM, OWASP Agent, MITRE ATLAS, jailbreak, system-prompt exfiltration, and custom playbooks. Strategies range from zero-shot and few-shot to multi-shot, iterative, multimodal, jailbreak, and chain-of-thought testing. That breadth helps teams assemble varied runtime checks, including prompt injection, jailbreak, data leakage, and custom cases; it does not replace static code review.

Evidence and workflow

Each run captures prompts, uploads, responses, and tool calls in attack_log.json. A separate pipeline_report.json summarizes categories and severity, giving reviewers a record of what the target received and returned. Results can also be exported through AIRTA Systems’ bulk import API for review and regression tracking.

The repository describes the project as deprecated and no longer maintained, and points users to Genbounty LLM Hunter instead. That is a substantial drawback for teams choosing a tool to anchor ongoing security work: its documented workflow may fit, but there is no active maintenance to rely on.

Pricing

AIRTA Red Team costs nothing. The open-source plan has no license fee or subscription, with no seat or quota cap stated in the plan. It does require users to provide their own LLM API key, so the software is free but test-generation and assessment access depends on that key.

There are no paid tiers to trade off against the free plan. This suits teams comfortable operating self-hosted open-source software and supplying their own API credentials; teams needing maintained support should look elsewhere.

Platforms

AIRTA Red Team supports API, Linux, macOS, web, and Windows, with self-hosted deployment. Docker and CI are among the described ways to run it, allowing teams to place assessments within their own environment or direct them at authorized external endpoints.

Who it's for

Red teams, whitehat testers, pentesters, and AppSec or MLsec teams are the clearest fit, particularly when they need runtime checks against LLM applications and evidence of prompts, uploads, responses, and tool calls. It is less suitable for teams seeking source-code analysis, an actively maintained project, or a workflow that does not depend on a user-supplied LLM API key.

Pros and cons

• Pro: Multiple playbooks and attack strategies give teams a broad basis for authorized runtime assessments.
• Pro: Browser and direct HTTP testing, including file uploads and tool calls, cover varied target interactions.
• Pro: Captured run logs and severity rollups provide concrete evidence for review and regression tracking.
• Pro: The open-source plan carries no license fee or subscription.
• Con: The project is deprecated and no longer maintained, making it a risky foundation for continuing security programs.
• Con: Users must supply an LLM API key, so the free license does not remove the need for API access.
• Con: Its assessment scope is runtime behavior, not source-code static analysis.

Alternatives

AI Security Testing Tools is the broader category to explore when comparing approaches. Consider OpenSecureAI Scanner if a freemium option with a published Pro tier is preferable: Pro costs $49 a month and includes 50,000 API requests monthly, 120 requests per minute per key, hosted AI remediation and LLM Gateway completions, and priority email.

Project Moonshot is another free option for teams comparing AI security tools. Promptfoo may suit teams wanting a published free Community plan with 10k red-team probes per month, LLM evaluation features, model providers and integrations, and local or self-hosted runs. For a scan-based alternative, PromptGuard has a free tier with 20,000 scans a month, one API key, one project, and 24-hour log retention.

PyRIT is a free open-source framework for teams prepared to use a Python environment and configure AI endpoints. Augustus is a free Apache 2.0 option for Linux, macOS, and Windows, with provider credentials potentially needed for API usage. Basilisk offers free AGPL-3.0 software distributed as CLI, desktop, Docker, or source, for authorized use. AgentScan may be a better fit for teams seeking a freemium scan service: its free plan allows five scans per month and includes 185 attack vectors, five categories, agent-format detection, and JSON results.

Verdict

AIRTA Red Team is a reasonable fit for security teams that need a no-license-fee, self-hosted way to run authorized LLM runtime assessments and retain structured evidence. Its varied attack coverage and browser/API workflows are useful strengths. Choose another tool if ongoing maintenance is a requirement: the project is deprecated, and every user must also bring an LLM API key.