3 Commits

Author	SHA1	Message	Date
dayuan.jiang	46cbc3354c	fix: add manual token usage reporting to Langfuse for Bedrock streaming ... Bedrock streaming responses don't auto-report token usage to OpenTelemetry. This fix manually sets span attributes (ai.usage.promptTokens, gen_ai.usage.input_tokens) from the AI SDK onFinish callback to ensure Langfuse captures token counts.	2025-12-05 00:26:02 +09:00
dayuan.jiang	46d2d4e078	refactor: add input validation and singleton pattern for Langfuse API routes ... - Add Zod schema validation for log-feedback and log-save endpoints - Create singleton LangfuseClient to avoid per-request instantiation - Simplify log-save to only flag trace (no XML content sent) - Use generic error messages to prevent info leakage	2025-12-04 23:44:00 +09:00
Dayuan Jiang	9d9613a8d1	feat: add trace-level input/output to Langfuse observability (#69) ... * feat: add trace-level input/output to Langfuse observability - Add @langfuse/client and @langfuse/tracing dependencies - Wrap POST handler with observe() for proper tracing - Use updateActiveTrace() to set trace input, output, sessionId, userId - Filter Next.js HTTP spans in shouldExportSpan so AI SDK spans become root traces - Enable recordInputs/recordOutputs in experimental_telemetry * refactor: extract Langfuse logic to separate lib/langfuse.ts module	2025-12-04 11:24:26 +09:00