How much does ChatNexus TTS cost?
Costs vary with response length. Longer knowledge base snippets cost more to synthesize. OpenAI pricing page lists the exact rates by model and region.
How does ChatNexus caching work?
The cache is keyed by sha256(text), so identical texts always hit the same cache entry. This is especially effective for common fallback messages, greetings, and pre-synthesized 'thinking' sounds. Cache hits save the full OpenAI API call cost.
What's the typical cost savings from caching?
A support bot answering the same 50 FAQs all day will hit cache on ~60% of requests. A research bot generating unique queries will see lower rates. ChatNexus provides cost analytics in the dashboard so you can monitor your savings.
Do I need to set up caching myself?
The cache is transparent. ChatNexus manages key generation, TTL, and eviction. You can view cache hit rates in your agent's analytics dashboard under the TTS Costs section.
What are pre-synthesized filler phrases and how do they save money?
When an agent is processing a knowledge base query, ChatNexus can play a pre-recorded 'thinking' sound while waiting. These fillers are cached globally and reused millions of times, eliminating redundant TTS API calls across the entire platform.