Import sessions from a Langfuse LangGraph trace export
Imports production traces from a Langfuse JSON export targeting a LangGraph application.
The file is uploaded to storage and a background BullMQ job processes it; the response is the
jobId you can poll on GET /jobs/{jobId}/status. All content validation (Langfuse shape,
LangGraph framework signal, drop rules) happens inside the job, not at the request boundary —
a syntactically valid upload that fails semantic checks is still accepted with 202 and the
rejection surfaces as failed state with a message in the job status error field.
Two request shapes are supported:
multipart/form-datawith a singlefilefield plusversionId.application/jsonwith atextfield carrying the JSON content as a string, plusversionIdand optionaloriginalFilename.
Exactly one of file (multipart) or text (JSON) must be present.
isProduction is optional and defaults to true, preserving the legacy
always-production behaviour for callers that have not been updated.
Multipart clients may send "true"/"false" as a string because form fields
are inherently strings; JSON clients must send a native boolean.
Documentation Index
Fetch the complete documentation index at: https://docs.galtea.ai/llms.txt
Use this file to discover all available pages before exploring further.
Authorizations
API key authorization. Pass your API key in the Authorization header as a Bearer token. Both new (gsk_*) and legacy (gsk-) API keys are accepted, e.g. Authorization: Bearer gsk_... or Authorization: Bearer gsk-....
Body
The Langfuse trace export (JSON). Max 25MB.
Id of the Version the imported sessions belong to.
Whether the imported sessions are production traffic ("true") or
development traffic ("false"). Defaults to "true" when omitted.
true, false Response
Import accepted; poll the returned jobId for completion.
Id of the BullMQ job processing the import. Poll /jobs/{jobId}/status for state.
"job_abc123"