Settings
Configure your Cascabel experience
When disabled, all AI insight cards are hidden across the app.
Using /
GPU-enabled hardware recommended. CPU inference may result in slow response times for interactive features.
Some features require structured JSON output (e.g. Insights Summary). Verify your Ollama model supports this before use.
Maximum response length (100-32000)
Latency: ms
No AI usage data for this period
Provider Breakdown
| Provider / Model | Calls | Cost (USD) | Tokens |
|---|---|---|---|
| / |
Cost by Feature
Cost by Provider
| Time | Feature | Provider / Model | Tokens | Cost | Latency |
|---|---|---|---|---|---|
| / |
Allows T4 Advisory to fetch real-world benchmarks from the web
Reduce spacing in tables
Default period for Analysis page
Show floating chat button on all pages