.svg){kind=logo}
Multilingual & Multimodal Attacks:
How they work and how Lakera stops them
detected per day
How the attack works
A malicious user supplies a prompt containing a multilingual attack intended to extract hidden API keys.
The goal is to confuse and exploit models where refusal heuristics are not confidently trained across all languages.
This LLM understands the request enough to follow the instructions not enough to not classify the request as harmful
{
"data": {
"name": "AI Policy",
"policy_mode": "IO",
"input_detectors": [
{
"type": "prompt_attack",
"threshold": "l2_very_likely"
},
],
"output_detectors": [
{
"type": "pii/credit_card",
"threshold": "l2_very_likely"
},
{
"type": "pii/api_keys",
"threshold": "l2_very_likely"
}
],
"id": "policy-9b52e331-d609-4ce3-bbb9-d2b1e72a0f20"
}
}
Lakera Guard’s integration understands 100+ languages
Our Prompt Defense guardrails detect the data exfiltration attempt when checking the input prompt.
Our customizable Data Leakage Prevention guardrails will detect, log (and redact) sensitive data that may elude LLM guardrails for novel prompt attacks
Lakera blocks unsafe instructions, detects disguised intent, redacts any sensitive entities (names, salaries), and logs the event for audit and review.
{
"payload": [],
"flagged": true,
"dev_info": {
"timestamp": "2025-11-24T12:35:12Z",
},
"metadata": {
"request_uuid": "ce8180b1-26bc-4177-9d7f-54ca7377378a"
},
"breakdown": [
{
"project_id": "project-7539648934",
"policy_id": "policy-a2412e48-42eb-4e39-b6d8-8591171d48f2",
"detector_id": "detector-lakera-default-prompt-attack",
"detector_type": "prompt_attack",
"detected": true,
"message_id": 0
}
]
}
How Lakera Stops Link-based Prompt Attacks
.svg)
Catch instruction overrides, jailbreaks, indirect injections, and obfuscated prompts as they happen, before they reach your model.
Block, redact, or warn. Fine-tune with allow-lists and per-project policies to minimize false positives without weakening protection.
.svg)
Lakera Guard continuously learns from 100K+ new adversarial samples each day. Adaptive calibration keeps false positives exceptionally low.
Protects across 100+ languages and evolving multimodal patterns, with ongoing support for image and audio contexts.
.svg)
Full audit logging, SIEM integrations, and flexible deployment options, SaaS or self-hosted, built for production-scale GenAI systems.
Works seamlessly with enterprise environments
Frequently asked questions
Yes. Lakera supports multilingual threat detection (including cross-language attacks) and handles indirect or obfuscated instructions embedded in text or mixed languages.
Yes. You can set a policy’s flagging sensitivity (L1 … L4) and apply that policy to different projects or use-cases (including multilingual/multimodal).
Detections across 100 + languages are logged in the dashboard and request logs, allowing you to filter by language, project, threat type and time-range.
Deploy AI with confidence
Related attack patterns
