PromptGuard
promptguard.co
- Free plan
- Windows
- macOS
- Linux
- Web
- Browser extension
- API
- Self-hosted
What PromptGuard is
PromptGuard is a security layer between an application and its LLM provider that checks requests before they reach the model. It describes 15 detectors across six layers for threats such as prompt injection, jailbreaks, PII, data exfiltration, toxicity, fraud, secrets, malware and tool injection. The product says it detects and redacts 43 PII types with reversible tokenization. Its SDK can automatically instrument supported LLM calls with one initialization call while leaving existing provider code unchanged. Listed providers include OpenAI, Anthropic, Gemini, Azure, Bedrock, Mistral, Cohere, DeepSeek, Groq, HuggingFace, Ollama and vLLM. Deployment choices include managed cloud, hybrid self-hosting and air-gapped deployment. Zero-retention mode does not store prompt or response content; default security events retain a shortened preview and content hash. PromptGuard says customer content is not used to train, fine-tune or evaluate models. The permanent Free tier includes 20,000 scans a month, one API key, one project and 24-hour log retention. Paid plans include a 14-day money-back guarantee, not a trial.
Who it suits
It suits teams adding request inspection and PII controls to applications that use LLM providers. Its deployment options include managed cloud, hybrid self-hosting and air-gapped use.
Good to know, and what to watch
- Describes 15 detectors across six layers.
- Detects and redacts 43 PII types.
- Supports multiple named LLM providers.
- Offers managed cloud, hybrid self-hosting and air-gapped deployment.
- Free tier includes 20,000 monthly scans.
- No time-limited free trial is offered.
- Hosted service does not currently offer an EU region.
- Five OWASP LLM Top 10 risks are only partially covered.
Plans
Every plan and priceEnterprise
Scale
Free
Team
Free plan
Getting started
- Open the PromptGuard website
- Choose managed cloud, hybrid self-hosting, or air-gapped deployment
- Select a plan and create an account or project
- Connect a supported model provider
- Initialize the SDK to instrument supported calls
- Configure detectors, PII handling, and retention settings
Limits to check
The Free tier allows 20,000 scans monthly, one API key, one project, and 24-hour log retention. Team provides 15,000 pooled scans per seat, while Pro provides 100,000 scans monthly; Scale includes 500,000 requests monthly with metered overage.
The Club's review
Overview
PromptGuard screens application requests before they reach an LLM provider, for teams that need security controls around model traffic. It combines broad threat detection with privacy controls and multiple deployment modes, but its hosted service is US-based and its stated OWASP coverage has gaps.
Key features
Threat detection and testing
PromptGuard describes 15 detectors across six layers for risks including prompt injection, jailbreaks, PII, data exfiltration, toxicity, fraud, secrets, malware and tool injection. It also supports tests for prompt injection, jailbreaks, data leakage, unsafe outputs and custom cases. That mix suits teams looking to screen live requests and assess application behavior; it is not a complete answer to LLM application security. PromptGuard says it covers five OWASP LLM Top 10 risks in full and five partially, with provenance verification and vector-store isolation among the gaps.
Privacy, integrations and deployment
The product says it detects and redacts 43 PII types, with reversible tokenization. This can mask sensitive values while preserving the option to restore them. Its SDK can instrument supported LLM calls with one initialization call and leave existing provider code unchanged, reducing the integration work for teams already using supported providers. Those include OpenAI, Anthropic, Gemini, Azure, Bedrock, Mistral, Cohere, DeepSeek, Groq, HuggingFace, Ollama and vLLM.
Managed cloud, hybrid self-hosting and air-gapped deployment give organizations options for different infrastructure and control requirements. Air-gapped licences validate offline with a signed key. In zero-retention mode, prompt and response content is not stored; default security events instead retain a truncated 500-character preview and content hash. PromptGuard says customer prompts, completions and documents are never used to train, fine-tune or evaluate models.
The hosted service runs on Google Cloud Run in us-central1, and the company says it does not offer a hosted EU region. Organizations that require EU-hosted service should weigh that limitation or consider self-hosting. GDPR and CCPA are supported; SOC 2 Type II certification has not yet been achieved, and ISO 27001 is planned.
Pricing
PromptGuard has a permanent free tier rather than a time-limited trial. Paid plans include a 14-day money-back guarantee, and no free trial is offered.
| Plan | Price | What it includes |
|---|---|---|
| Free | Free | 20,000 scans a month, 1 API key, 1 project and 24-hour log retention; community support. |
| Team | $19 a month | 15,000 scans per seat, pooled; browser extension, macOS and Windows agent, fleet enrollment, MDM and org-wide policy; email support with a 48-hour response time. |
| Pro | $99 a month | 100,000 scans a month, 5 API keys, 5 projects and 7-day log retention. |
| Scale | Custom pricing | 500,000 requests a month, unlimited API keys and projects, 30-day log retention, and overage metered at $0.40 per 1,000 requests up to a spend cap; priority support with a 24-hour response time. |
| Enterprise | Custom pricing | Custom volume, fully air-gapped deployment, SCIM, IP allowlist, custom retention and dedicated support with a four-hour response SLA. |
Free is a useful starting point for a small project, but the single key and project, low retention and community support constrain team use. Team adds desktop and browser tooling, fleet management and pooled per-seat scans; its 15,000-scan allowance is a meaningful limit to check against expected traffic. Pro raises the monthly quota and key, project and retention caps, but costs more and does not state the Team plan's fleet features. Scale is aimed at higher-volume use, with an explicit overage rate and spend cap; Enterprise is the fit for custom volumes and air-gapped deployments. Both use custom pricing.
Platforms
PromptGuard supports API, browser extension, Linux, macOS, self-hosted, web and Windows. Its deployment options span managed cloud, hybrid self-hosting and air-gapped installations, making it more flexible than a hosted-only gateway for organizations with strict infrastructure constraints.
Who it's for
PromptGuard is best suited to teams adding inspection and data controls to applications that use multiple LLM providers, especially where self-hosting or offline deployment matters. Smaller projects can start on Free, while teams that need organizational policy and endpoint tooling have a more relevant fit in Team. It is a weaker choice for organizations that require a hosted EU region or complete OWASP LLM Top 10 coverage.
Pros and cons
Pros
- Flexible deployment: Managed cloud, hybrid self-hosting and fully air-gapped deployment cover varied infrastructure requirements.
- Broad provider support: The named integrations range from major cloud providers to Ollama and vLLM, useful for teams working across hosted and local models.
- Practical privacy controls: Reversible PII tokenization and zero-retention mode address sensitive data handling, while default event previews are limited to 500 characters.
- Low-cost entry point: The permanent free tier includes 20,000 monthly scans, although its single project and API key limit its scope.
Cons
- Hosted region constraint: The hosted service runs in us-central1, with no hosted EU region.
- Incomplete stated OWASP coverage: Five risks are only partially covered, and provenance verification and vector-store isolation are named gaps.
- Certification status: SOC 2 Type II is not yet certified, while ISO 27001 remains planned.
- Plan limits need scrutiny: Free has only 24-hour log retention, and Team's scans are capped per seat even though pooled.
Alternatives
For an LLM security shortlist, compare LLM Security Tools, AI Guardrail Software and AI Security Testing Tools.
- Amazon Bedrock Guardrails is a paid, web-based option with separately priced text content filters and denied topics, and image content filters listed as free. Choose it when those Bedrock guardrail controls and their listed pricing fit better than PromptGuard's broader deployment choices.
- GuardionAI is a paid alternative with a free trial and an Enterprise plan billed by custom contract, offering unlimited requests and seats with up to 365-day log retention. Consider it if a trial and longer retention matter more than PromptGuard's published entry-level plans.
- Openlayer Guardrails has a free Basic plan with one member, five projects, 20,000 inference logs per month, 20 tests per project and three months of data retention. It is worth considering when those testing and retention limits suit your needs and PromptGuard's threat-inspection deployment options are not essential.
- GLACIS offers a free account with no card required and setup described as taking about a minute, for readers who prioritize a simple free entry point.
- WitnessAI is a paid API, web and Windows option with custom Enterprise pricing.
- Lasso AI Security Platform is a paid option for API, extension and web platforms.
- Prisma AIRS AI Gateway is a paid API, self-hosted and web option with consumption-based licensing through Software NGFW credits metered by token usage. Consider it when token-based gateway licensing better fits your procurement model.
- HiddenLayer AI Guardrails is a paid, self-hosted and web option priced at $5,000,000 a year on a 12-month contract; additional AWS infrastructure costs may apply. It is a different price and contract commitment from PromptGuard's free and per-month entry plans.
Verdict
Choose PromptGuard if your team needs request inspection, reversible PII controls and deployment flexibility across cloud, self-hosted or air-gapped environments. Its strongest case is combining provider breadth with those data controls; look elsewhere if hosted EU residency or more complete OWASP LLM Top 10 coverage is a requirement.
The verdict
PromptGuard offers request inspection, data controls and several deployment modes for LLM applications. Review its stated coverage gaps and hosted-region limits against your requirements.
Questions people ask
Is PromptGuard free?
Yes. The permanent Free plan costs nothing and includes 20,000 scans a month, one API key, one project, and 24-hour log retention.
Is there a free trial?
No. PromptGuard describes Free as a permanent tier; paid plans include a 14-day money-back guarantee.
Which providers are supported?
The listed providers include OpenAI, Anthropic, Gemini, Azure, Bedrock, Mistral, Cohere, DeepSeek, Groq, HuggingFace, Ollama, and vLLM.
Can PromptGuard run without hosted cloud?
Yes. Deployment choices include hybrid self-hosting and air-gapped deployment, as well as managed cloud.
What happens to prompts and responses?
Zero-retention mode stores no prompt or response content. Default security events record a truncated 500-character preview and a content hash.
What does the Team plan cost?
Team is listed at $19 a month and includes 15,000 pooled scans per seat, a browser extension, macOS and Windows agents, fleet enrollment, MDM, and organization-wide policy.
What the maker publishes
- AI-agent security
- Agent security checks tool calls, privilege escalation and agent behavior before actions execute. source
- Apple app pricing
- The Apple listing is free with in-app purchases priced at $9.99 monthly or $79.99 annually for Pro, and $14.99 monthly or $114.99 annually for Family. source
- Apple app requirements
- The Apple listing requires iOS or iPadOS 16, and macOS 13 on an Apple M1-or-later Mac. source
- Benchmark F1
- PromptGuard reports an F1 score of 0.887 on 2,369 samples across eight datasets. source
- Benchmark precision
- The published benchmark reports 99.1% precision, with 10 false positives among 991 benign samples. source
- Benchmark recall
- The published benchmark reports 80.3% recall and says 271 of 1,378 attack samples were missed. source
- Certifications
- The trust page says SOC 2 Type II is not yet certified, ISO 27001 is planned, and GDPR and CCPA are supported. source
- Chrome extension features
- The listing describes live warnings, PII detection, one-click redaction, history scanning, a local dashboard, allowlists and custom keywords. source
- Chrome extension privacy
- The Chrome listing says detection runs locally and prompt or conversation text is not sent to the developer's servers. source
- Company
- PromptGuard is the trading name of PG Tech Ltd, registered in England and Wales, with a registered office in London. source
- Company size and headquarters
- LinkedIn lists PromptGuard as privately held, headquartered in London, with 2–10 employees. source
- Data retention
- Zero-retention mode avoids writing prompt and response bodies; metadata-only logs retain decisions, reasons and timestamps. source
- Deployment
- The product offers managed cloud, hybrid self-hosting, and air-gapped deployment; air-gapped licences validate offline with a signed key. source
- Deployment modes
- Customers can use managed cloud, hybrid self-hosting in their VPC, or a fully air-gapped installation. source
- Documentation coverage
- The documentation covers quickstart, APIs, SDKs, streaming, providers, CLI, VS Code, Cursor, GitHub scanning, MCP, dashboards, webhooks, audit logs, SSO, SCIM, self-hosting and reliability. source
- Failure behavior
- The SDK defaults to fail-open when the Guard API is unavailable, with an option to fail-closed. source
- Hosted region
- The hosted service runs on Google Cloud Run in us-central1 with managed Postgres, and no hosted EU region is offered. source
- Integration methods
- Developers can use auto-instrumentation, a direct Guard API, an HTTP proxy or a GitHub scanner. source
- Integrations
- The site lists integrations/providers including OpenAI, Anthropic, Gemini, Azure, Bedrock, Mistral, Cohere, DeepSeek, Groq, HuggingFace, Ollama, and vLLM. source
- Latency
- Deterministic fast-path checks are advertised as taking under 10 milliseconds, while escalated checks are network-bound. source
- Legal maker
- PromptGuard identifies PG Tech Ltd as its England-and-Wales trading company, registered at 128 City Road, London, EC1V 2NX. source
- Limit: G2 enterprise depth
- The sole G2 reviewer asks for more enterprise-grade capabilities, and G2 says there are too few reviews for buying insight. source
- Limit: independent review false positives
- AI Indigo reports false positives on complex legitimate prompts and says sensitivity tuning may be required. source
- Limit: independent review latency
- AI Indigo measured roughly 0.15 seconds of added delay and warns that this could matter for ultra-low-latency or voice applications. source
Checked against the maker's pages on 14 Oct 2026.