@anvia/langfuse API reference

All public symbols are exported from @anvia/langfuse.

Tracing and scoring

const langfuse: {
  create(options?: LangfuseTracingOptions): LangfuseTracing
}

type LangfuseTracing = AgentObserver & {
  flush(): Promise<void>
  shutdown(): Promise<void>
  score(args: LangfuseScoreArgs): Promise<void>
  flushScores(): Promise<void>
  scoreQueueDepth(): number
  getCurrentTrace(): LangfuseTraceHandle | undefined
}

LangfuseTracingOptions configures credentials, baseUrl, environment and release names, service identity, timeouts, score batching/retries, safe or full capture, byte limits, and input/output redaction.

type LangfuseScoreArgs = {
  traceId?: string
  observationId?: string
  name: string
  value: number | string
  dataType?: 'NUMERIC' | 'CATEGORICAL' | 'BOOLEAN'
  comment?: string
  metadata?: Record<string, JsonValue | undefined>
  configId?: string
  scoreConfigId?: string
  environment?: string
  timestamp?: Date | string
}

LangfuseScoreError contains the rejected scores and original cause when available.

Evaluation reporting

function createLangfuseEvalReporter<Input = unknown, Output = unknown, Expected = unknown>(
  tracing: Pick<LangfuseTracing, 'score'>,
  options?: LangfuseEvalReporterOptions,
): EvalReporter<Input, Output, Expected>

Reporter options control invalid results, strict failure, missing traces, input truncation, messages, and context.

function runEvalAsExperiment<Input, Output, Expected = unknown>(
  evalOptions: RunEvalSuiteOptions<Input, Output, Expected>,
  experimentOptions: RunEvalAsExperimentOptions<Input, Output, Expected>,
): Promise<RunEvalAsExperimentResult<Input, Output, Expected>>

The experiment options add Langfuse tracing, optional dataset client configuration, pagination, timeouts, score publishing, reporter configuration, and context inclusion. The result contains both the Anvia suite result and Langfuse dataset-run result.

Datasets

function createLangfuseDatasetClient(
  tracing: Pick<LangfuseTracing, 'score'>,
  options?: LangfuseDatasetClientOptions,
): LangfuseDatasetClient

interface LangfuseDatasetClient {
  createDataset(input: {
    name: string
    description?: string
    metadata?: Record<string, JsonValue | undefined>
  }): Promise<LangfuseDataset<unknown, unknown>>
  getDataset<Input, Expected>(name: string): Promise<LangfuseDataset<Input, Expected>>
  upsertItems<Input, Expected>(
    name: string,
    items: LangfuseDatasetItem<Input, Expected>[],
  ): Promise<void>
  runExperiment<Input, Output, Expected>(
    options: LangfuseRunExperimentOptions<Input, Output, Expected>,
  ): Promise<LangfuseRunExperimentResult>
}

Dataset items contain id, typed input, optional expected, and JSON metadata. A run experiment invokes the supplied run(item) callback and returns posted count plus per-item errors.

Prompts

function createLangfusePromptClient(
  tracing: Pick<LangfuseTracing, 'score'>,
  options?: LangfusePromptClientOptions,
): LangfusePromptClient

interface LangfusePromptClient {
  getPrompt(name: string, options?: LangfusePromptGetOptions): Promise<LangfusePrompt>
  getPromptText(name: string, options?: LangfusePromptGetOptions): Promise<string>
  getPromptChat(name: string, options?: LangfusePromptGetOptions): Promise<LangfuseChatMessage[]>
  refresh(): void
}

LangfusePromptGetOptions selects a version or label, overrides cache TTL, or forces refresh. LangfusePrompt reports name, version, labels, text or chat content, type, optional tags, and resolution time.

Redaction

function createPiiRedactor(options?: LangfuseRedactionOptions): PiiRedactor

interface PiiRedactor {
  redactString(input: string): string
  redactObject<T>(input: T): T
  redactMessages(input: Message[]): Message[]
  patternNames(): string[]
}

DEFAULT_PATTERNS is the exported built-in pattern list. RedactorPattern has a name and RegExp; options can replace the built-ins and configure replacement text.

Export inventory

Area	Public exports
Tracing	langfuse, LangfuseTracing, LangfuseTracingOptions, LangfuseTraceHandle, LangfuseCaptureMode
Scoring	LangfuseScoreArgs, LangfuseScoreDataType, LangfuseScoreError
Evaluations	createLangfuseEvalReporter, LangfuseEvalReporterOptions, runEvalAsExperiment, RunEvalAsExperimentOptions, RunEvalAsExperimentResult
Datasets	createLangfuseDatasetClient, LangfuseDatasetClient, LangfuseDatasetClientOptions, LangfuseDataset, LangfuseDatasetItem, LangfuseRunExperimentOptions, LangfuseRunExperimentResult, LangfuseRunItemResult, LangfuseRunItemError
Prompts	createLangfusePromptClient, LangfusePromptClient, LangfusePromptClientOptions, LangfusePromptGetOptions, LangfusePrompt, LangfuseChatMessage
Redaction	createPiiRedactor, DEFAULT_PATTERNS, PiiRedactor, RedactorPattern, LangfuseRedactionOptions, LangfuseRedactionMode