Prompt caching
LiveStable prompt prefixes are cached. Faster responses, lower bill, passed to you.
- ✓Stable-prefix caching against Anthropic's prompt cache
- ✓Cached vs uncached breakdown in usage log
- ✓Pass-through pricing — you see the saved cost
- ✓5-minute cache TTL with automatic refresh
What it is.
We cache the stable prefix of our prompts (system prompts, Indian-law context, firm style) against Anthropic's prompt cache. Cache hits cost roughly 10% of a fresh call and respond in roughly half the time.
We pass the savings to you transparently — your usage logs show cached vs uncached tokens and the resulting cost.
Three steps.
End to end.
Cache miss — full cost, full latency.
Cache hit on the stable prefix — ~10% cost, ~half latency.
Cached vs uncached tokens broken out per request, so you can verify the savings.
What you get.
- ✓Stable-prefix caching against Anthropic's prompt cache
- ✓Cached vs uncached breakdown in usage log
- ✓Pass-through pricing — you see the saved cost
- ✓5-minute cache TTL with automatic refresh
Quick answers.
No — only the stable prefix (system prompts, model instructions, firm style). Your document content is never cached.
More in AI Controls & Guardrails.
We do not train models on customer content without explicit opt-in. Default is off.
User-pasted text is sanitised before reaching the LLM.
Daily and monthly cost ceilings per organisation. Soft warnings, hard cutoffs.
200K input / 8K output limits with graceful truncation and a clear notice.