Built-in AI brings zero marginal cost, no API keys, and permissionless innovation. But how do you debug and monitor quality of client side AI? Julien Bataillé (Rakuten) built a PoC instrumenting the Prompt API with OpenInference spans & direct OTLP export: github.com/rakutenanaly...
github.comweb-ai-demos/prompt-api-telemetry at prompt-api-observability · rakutenanalytics/web-ai-demosContribute to rakutenanalytics/web-ai-demos development by creating an account on GitHub.
0 reposts2 likes