OS

AI Red Teaming Tools

OpenSecureAI Scanner: AI Security Checks and Pricing

opensecureai.com · checked 15 Oct 2026

Open-source AI security scanner with local tools, hosted detections, and model gateway checks.

How it plugs in

YesPublic API
YesOpen source
YesFree tier

Signal on AI Red Teaming Tools7.7#2 of 27

OpenSecureAI Scanner is an AI security platform for finding prompt injection, jailbreaks, leaked data, and related threats. Its scanner runs locally through a browser, CLI, GitHub Action, or npm package, while authenticated API calls add hosted detections. It also offers prompt analysis, firewall rules, agent checks, observability, and a gateway for OpenAI, Anthropic, xAI, and Groq.

OpenSecureAI Scanner home page

Review

In a line

OpenSecureAI Scanner is an open-source scanner and wider security platform for AI and generative-AI systems. It looks for prompt injection, data poisoning, model theft, and other related risks.

What it does

Untrusted text is checked for prompt-injection, jailbreak, system-prompt-leak, and data-exfiltration patterns. Through scanPrompt, the library reports a 0–100 score, a risk level, and findings tied to individual rules. It also supports prompt-injection, jailbreak, data-leakage, and unsafe-output testing.

The browser-based Unified Analysis workflow combines prompt-injection checks, secrets and personal-data scanning, and optional system-prompt auditing. It can identify API keys, tokens, private keys, email addresses, Social Security numbers, and payment cards, then show masked results.

Prompt Firewall can allow, flag, or block traffic. LLM Gateway scans an input, sends it to a chosen model provider, checks the response, and can mask leaked secrets or personal data or block risky traffic. Agent Guard reviews OpenAI, Anthropic, and MCP tool definitions for risky access and cross-tool exfiltration.

How it connects

Developers can use the scanner through npm, a command-line interface, a GitHub Action, or the REST API. The public API includes scan, unified-analysis, agent-guard, gateway, events, and threat-feed endpoints. Public TypeScript packages also cover the firewall and agent guard.

The basic engine runs locally in a browser, CLI, or GitHub Action. With authentication, the hosted service adds private checks for obfuscation and encoding evasions, indirect injection, tool abuse, multilingual overrides, and de-obfuscation. The core engine and npm packages are open source, but no specific license is named.

The gateway supports OpenAI, Anthropic, xAI, and Groq with customer-supplied keys. Authenticated scans save a dashboard summary, while raw prompt text is never stored; gateway provider keys are also never stored or logged.

Good fit

OpenSecureAI identifies AI application teams, agent and MCP developers, security and compliance reviewers, learners, and red-teamers as audiences. It can help teams scan prompts in development, gate pull requests, inspect agent tools, and review model traffic.

The browser playground includes five local capture-the-flag challenges covering direct overrides, filter bypasses, role-play jailbreaks, indirect injection, and encoding smuggling. LLM Observability groups injection attempts, blocks, and anomalies from API events; users are told to send labels and metadata instead of raw prompts or responses.

Watch for

Heuristic scans can produce false positives or miss issues, and the terms say they are not a security guarantee. Treat results as checks to review rather than proof that a system is safe.

The scanner package targets Node.js 18 or newer, uses ESM and TypeScript types, and has zero runtime dependencies. Hosted advanced detections require authenticated API access. Paid plans are not available through self-serve checkout; customers should contact sales for setup.

OpenSecureAI says its hosted advanced engine contains private detections, while its core engine and packages are open source. The pages do not name an open-source license.

Pricing

The Free plan costs $0 per month. It provides 2,000 API requests monthly, 30 requests per minute per key, browser tools and npm packages, the hosted advanced engine, scan history, LLM Gateway BYOK, and BYOK AI remediation.

Pro is priced at $49 per month. Its allowance is 50,000 API requests monthly with 120 requests per minute per key, plus hosted AI remediation, hosted LLM Gateway completions, and priority email support.

Team pricing is available on request and is billed as Custom. The plan lists 250,000 API requests per month and 300 requests per minute per key; team seats and SSO, custom rules, and an SLA are planned.

Verdict

Choose OpenSecureAI Scanner if you need an open-source starting point for prompt and AI-system checks, with the option to add hosted detections and model routing. Its local tools, pull-request action, REST endpoints, and provider gateway cover both development and operations.

It is less suitable as a sole security control because the scans are heuristic and can miss problems. Teams that need paid features must arrange setup with sales, and buyers should confirm whether the planned Team features or an unnamed open-source license meet their requirements. For other options, see AI Security Testing Tools or AI Red Teaming Tools.

Our take

OpenSecureAI Scanner is a reasonable fit for teams that want local prompt checks plus optional hosted analysis. The free plan gives access to browser tools, packages, API requests, and gateway features, while paid plans add higher limits and hosted services. Its broad toolset covers prompts, secrets, agent definitions, and model traffic. However, heuristic results can be wrong or incomplete, and the product is not a security guarantee. Check the Node.js requirement, hosted-engine needs, planned Team items, and licensing details before adopting it for a critical workflow.

  • Open-source core engine and npm packages
  • Runs locally in a browser, CLI, or GitHub Action
  • Checks prompts, secrets, PII, agents, and model traffic
  • Supports OpenAI, Anthropic, xAI, and Groq
  • Free tier provides 2,000 API requests each month
  • Heuristic scans can miss issues or raise false positives
  • Paid checkout requires contacting sales
  • Hosted advanced detections need authenticated API access
  • No specific open-source license is named
  • Node.js 18 or newer is required for the scanner package
  • Pick it ifAI teams adding prompt and agent security checks; Developers needing local scans or pull-request gates; Security reviewers and red-team learners
  • Skip it ifYou need guaranteed detection rather than heuristic checks; You require self-serve paid checkout; You need a named open-source license before adoption

Pricing

The no-cost tier includes 2,000 monthly API requests. Pro is $49/month with 50,000 requests and added hosted services. Team pricing is custom for 250,000 requests, and paid setup is arranged through sales.

  • Free$0/month · 2,000 API requests/month · 30 requests/minute per key · browser tools and npm packages · hosted advanced engine · scan history · LLM Gateway BYOK · BYOK AI remediationFree
  • Pro$49/month · 50,000 API requests/month · 120 requests/minute per key · hosted AI remediation · hosted LLM Gateway completions · priority email support$49/mo
  • TeamCustom · 250,000 API requests/month · 300 requests/minute per key · team seats and SSO planned · custom rules and SLA plannedOn request

As opensecureai.com lists them, checked 15 Oct 2026.

Questions

Is OpenSecureAI Scanner free?

Yes. The Free plan costs $0 per month and provides 2,000 API requests monthly, browser tools, npm packages, scan history, and gateway features.

What does the scanner detect?

It detects prompt injection, jailbreak, system-prompt-leak, and data-exfiltration patterns in untrusted text. Tests also cover data leakage and unsafe output.

Can it run without sending prompts to a server?

The basic engine runs locally in a browser, CLI, or GitHub Action. Authenticated API keys add hosted detections.

Which model providers work with the gateway?

The gateway names OpenAI, Anthropic, xAI, and Groq as supported providers and accepts customer-supplied keys.

Does OpenSecureAI store raw prompts?

Authenticated scans save a summary to the dashboard, but raw prompt text is never stored. Gateway provider keys are never stored or logged.

What does Pro cost?

Pro is priced at $49 per month. It provides 50,000 API requests monthly, hosted AI remediation, hosted gateway completions, and priority email support.

Is the code open source?

OpenSecureAI says its core engine and npm packages are open source. Its pages do not name a specific open-source license.