ta
TanStack
GitHub
ai
Workspace
GitHub
CI Pipeline Executions
Filtered
Runs
Demo
Insights
Compare tasks
Analytics
Sign in
Toggle sidebar
Overview
⌘K
ai
Loading workspace stats
Loading workspace insights...
Statistics interval
7 days
30 days
Latest CI Pipeline Executions
Status
Fix filter
Filter
Fuzzy
Filter range
Sort by
Sort by
Start time
Sort ascending
Sort descending
GitHub
ai
Workspace
Succeeded
main
d98a598c fix(ai-bedrock): forward Converse cache token counts to TokenUsage (#1300) The Converse stream processor and the two structuredOutput paths copied only inputTokens, outputTokens and totalTokens out of metadata.usage. Bedrock reports cacheReadInputTokens and cacheWriteInputTokens beside them, and inputTokens counts only the uncached remainder, so a cached request looked like a near-zero-input call and the cache write and read costs had no field. buildConverseUsage() in converse/usage.ts now maps both counts to promptTokensDetails.cachedTokens / cacheWriteTokens, the same builder shape as ai-anthropic and openai-base, and omits the details object when Bedrock omits the counters. Docs say where the counts appear. Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
by Sunghoon Kang
S
Succeeded
main
d98a598c fix(ai-bedrock): forward Converse cache token counts to TokenUsage (#1300) The Converse stream processor and the two structuredOutput paths copied only inputTokens, outputTokens and totalTokens out of metadata.usage. Bedrock reports cacheReadInputTokens and cacheWriteInputTokens beside them, and inputTokens counts only the uncached remainder, so a cached request looked like a near-zero-input call and the cache write and read costs had no field. buildConverseUsage() in converse/usage.ts now maps both counts to promptTokensDetails.cachedTokens / cacheWriteTokens, the same builder shape as ai-anthropic and openai-base, and omits the details object when Bedrock omits the counters. Docs say where the counts appear. Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
by Sunghoon Kang
S
Succeeded
fix/bedrock-converse-cache-usage
4e5adbb9 fix(ai-bedrock): forward Converse cache token counts to TokenUsage The Converse stream processor and the two structuredOutput paths copied only inputTokens, outputTokens and totalTokens out of metadata.usage. Bedrock reports cacheReadInputTokens and cacheWriteInputTokens beside them, and inputTokens counts only the uncached remainder, so a cached request looked like a near-zero-input call and the cache write and read costs had no field. buildConverseUsage() in converse/usage.ts now maps both counts to promptTokensDetails.cachedTokens / cacheWriteTokens, the same builder shape as ai-anthropic and openai-base, and omits the details object when Bedrock omits the counters. Docs say where the counts appear. Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
by Sunghoon Kang
S
Succeeded
fix/bedrock-converse-cache-usage
4e5adbb9 fix(ai-bedrock): forward Converse cache token counts to TokenUsage The Converse stream processor and the two structuredOutput paths copied only inputTokens, outputTokens and totalTokens out of metadata.usage. Bedrock reports cacheReadInputTokens and cacheWriteInputTokens beside them, and inputTokens counts only the uncached remainder, so a cached request looked like a near-zero-input call and the cache write and read costs had no field. buildConverseUsage() in converse/usage.ts now maps both counts to promptTokensDetails.cachedTokens / cacheWriteTokens, the same builder shape as ai-anthropic and openai-base, and omits the details object when Bedrock omits the counters. Docs say where the counts appear. Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
by Sunghoon Kang
S
Succeeded
fix/bedrock-converse-cache-usage
4e5adbb9 fix(ai-bedrock): forward Converse cache token counts to TokenUsage The Converse stream processor and the two structuredOutput paths copied only inputTokens, outputTokens and totalTokens out of metadata.usage. Bedrock reports cacheReadInputTokens and cacheWriteInputTokens beside them, and inputTokens counts only the uncached remainder, so a cached request looked like a near-zero-input call and the cache write and read costs had no field. buildConverseUsage() in converse/usage.ts now maps both counts to promptTokensDetails.cachedTokens / cacheWriteTokens, the same builder shape as ai-anthropic and openai-base, and omits the details object when Bedrock omits the counters. Docs say where the counts appear. Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
by Sunghoon Kang
S
Succeeded
fix/bedrock-converse-cache-usage
4e5adbb9 Merge 7c9364a3d2a492bdf9b378e9d33cf787d512c8d8 into 416a4e34bb2728529a3cdc598e0469f356121732
by Sunghoon Kang
S
Previous
Next