Open-source analytics dashboard for AWS Bedrock. Track usage, cost, and performance across your teams and applications.
Seven dashboard views covering every aspect of your Bedrock usage
Real-time snapshot of your Bedrock activity - total invocations, token counts, cost, invocation charts, model distribution, and cost by team at a glance.

Everything you need to understand and optimize how your organization uses AWS Bedrock
Total invocations, tokens, and cost at a glance. Charts for invocation patterns, model distribution, and cost by team — plus the last 20 API calls.
See which teams and apps are active, what models they prefer, and how adoption grows. Identify underutilized capacity and champion teams.
Track input and output tokens by model, team, and application. Understand consumption patterns and input/output ratios across your organization.
Automatic cost calculation with per-model pricing. Break down spend by team, app, and model. High-cost request alerts and top spender identification.
Monitor latency percentiles (p50/p95/p99), error rates, and throughput. Status code breakdown, request size distribution, and token waste analysis.
Analyze usage over configurable date ranges. Daily invocations, cost, and token trends with day-by-day summaries and team cost breakdowns.
Filter by date range, model, team, and application across every page. Active filter chips with quick removal. Every data table exports to CSV.
Map IAM roles to human-readable team and app names. Override per-model pricing. Update config without redeploying — just upload to S3.
Your data never leaves your infrastructure. Deploy with SST in minutes.
From zero to dashboard in minutes
Turn on model invocation logging in the AWS Bedrock console. Point it at a CloudWatch Logs group.
Clone the repo, set basic auth secrets with npx sst secret set, and run npm run deploy.
Map IAM roles to team and app names in a simple YAML config. Upload it — no redeploy needed.
Free, open-source, and deploys to your own AWS account.