Prompt injection defence
LiveUser-pasted text is sanitised before reaching the LLM.
- ✓Pattern-based sanitisation on text inputs
- ✓Untrusted-input labelling to the LLM
- ✓Suspicious-pattern flag in the audit log
- ✓Continuous updates as new injection patterns emerge
What it is.
Prompt injection — where adversarial text in a document tries to hijack the AI ('ignore previous instructions and …') — is a real attack class. LexVio sanitises user-pasted text and uploaded document content through a hardened input layer before it reaches the LLM.
Document content is also clearly labelled to the model as 'untrusted input', so the model treats it as data, not instructions.
Three steps.
End to end.
Upload or paste. Text extraction kicks off.
Known injection patterns are neutralised. Suspicious patterns are flagged.
Document content is wrapped in clear delimiters and labelled as untrusted input. The model knows not to follow instructions inside it.
What you get.
- ✓Pattern-based sanitisation on text inputs
- ✓Untrusted-input labelling to the LLM
- ✓Suspicious-pattern flag in the audit log
- ✓Continuous updates as new injection patterns emerge
Quick answers.
No defence is perfect. We layer multiple controls and continuously update — but we don't make absolute claims.
More in AI Controls & Guardrails.
We do not train models on customer content without explicit opt-in. Default is off.
Daily and monthly cost ceilings per organisation. Soft warnings, hard cutoffs.
Stable prompt prefixes are cached. Faster responses, lower bill, passed to you.
200K input / 8K output limits with graceful truncation and a clear notice.