langfuse-mcp-java

Query Langfuse cost, token, latency, and usage analytics via the Metrics API. Mirrors: GET /api/public/metrics?query=

Pass the full query as a JSON string. All aggregation is server-side.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ QUERY STRUCTURE ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ { "view": string, // REQUIRED. traces | observations | scores-numeric | scores-categorical "metrics": [...], // REQUIRED. At least one { measure, aggregation, alias? } "fromTimestamp": string, // REQUIRED. ISO-8601 e.g. "2026-03-18T00:00:00Z" "toTimestamp": string, // REQUIRED. ISO-8601 e.g. "2026-03-25T23:59:59Z" "dimensions": [...], // Optional. [{ "field": "..." }] "filters": [...], // Optional. [{ "column", "operator", "value", "type", "key"? }] "timeDimension": {...}, // Optional. { "granularity": "hour|day|week|month|auto" } "config": {...} // Optional. { "bins": 10, "row_limit": 100 } }

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ VIEW ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ traces → end-to-end cost, tokens, latency per request observations → per LLM call; USE for model breakdowns (providedModelName) scores-numeric → numeric/boolean evaluation scores scores-categorical → categorical evaluation scores

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ MEASURES (by view) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ traces: count | observationsCount | scoresCount | latency | totalTokens | totalCost observations: count | latency | totalTokens | totalCost | timeToFirstToken | countScores scores-numeric: count | value scores-categorical: count ⚠ NEVER use inputTokens / outputTokens / promptTokens / completionTokens → 400 error. Use totalTokens.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ AGGREGATIONS ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ sum | avg | count | max | min | p50 | p75 | p90 | p95 | p99 sum → cost/token totals avg/p95/p99 → latency count → record counts

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ DIMENSIONS (group-by, by view) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ traces: name | tags | userId | sessionId | release | version | environment | observationName | scoreName observations: providedModelName | type | name | level | version | environment | userId | sessionId | traceName | traceRelease | traceVersion | promptName | promptVersion | scoreName ⚠ HIGH CARDINALITY — use in filters only, not dimensions: id | traceId | observationId

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ FILTERS ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Each filter: { "column", "operator", "value", "type", "key"? } type string/stringObject/boolean → operator: = | contains | does not contain | starts with | ends with type number/datetime → operator: = | < | > | <= | >= ⚠ NEVER use != / not_contains / not_equals for string fields → 400 error.

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ EXAMPLES ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Total cost last 7 days: {"view":"traces","metrics":[{"measure":"totalCost","aggregation":"sum"}], "fromTimestamp":"2026-03-18T00:00:00Z","toTimestamp":"2026-03-25T23:59:59Z"}

Daily cost trend this week: {"view":"traces","metrics":[{"measure":"totalCost","aggregation":"sum"},{"measure":"count","aggregation":"count"}], "timeDimension":{"granularity":"day"}, "fromTimestamp":"2026-03-18T00:00:00Z","toTimestamp":"2026-03-25T23:59:59Z"}

Cost by model: {"view":"observations","dimensions":[{"field":"providedModelName"}], "metrics":[{"measure":"totalCost","aggregation":"sum"},{"measure":"totalTokens","aggregation":"sum"}], "fromTimestamp":"2026-03-18T00:00:00Z","toTimestamp":"2026-03-25T23:59:59Z"}

Cost for a specific user: {"view":"traces","metrics":[{"measure":"totalCost","aggregation":"sum"}], "filters":[{"column":"userId","operator":"=","value":"user-123","type":"string"}], "fromTimestamp":"2026-03-18T00:00:00Z","toTimestamp":"2026-03-25T23:59:59Z"}

Production environment only: filters: [{"column":"environment","operator":"=","value":"production","type":"string"}]