Connect Langfuse
Forward Langfuse trace data to Obsivara by pointing a Langfuse webhook at the Obsivara ingest endpoint.
Configure the webhook
In your Langfuse project settings, navigate to Integrations → Webhooks and add a new webhook pointing to:
https://api.obsivara.com/ingest/webhook/langfuse?asset_id=my-langfuse-agentAdd the header:
Authorization: Bearer obs_live_YOUR_KEYLangfuse will POST a trace payload to Obsivara each time a trace completes. Obsivara maps the Langfuse trace id to a stable internal UUID automatically — you do not need to manage ID translation.
Accepted payload shape
Obsivara reads the following fields from the Langfuse webhook payload. All other fields are ignored.
| Field | Type | Description |
|---|---|---|
id | string | Langfuse trace ID (mapped to a stable UUID) |
name | string | Trace / pipeline name |
input | any | Input data passed to the trace |
output | any | Output data returned by the trace |
status | string | Trace outcome (success, error, etc.) |
input and output (singular), unlike LangSmith which uses inputs/outputs (plural). Make sure your webhook is pointing at /ingest/webhook/langfuse — not the LangSmith endpoint.Example payload:
{
"id": "trace_xyz789",
"name": "RAG pipeline",
"input": "What is the refund policy?",
"output": "You can request a refund within 30 days.",
"status": "success"
}Asset attribution
Pass ?asset_id= as a query parameter on the webhook URL to group traces under a named asset in the Obsivara dashboard. For example:
https://api.obsivara.com/ingest/webhook/langfuse?asset_id=rag-pipelineAll traces forwarded via this URL will appear under the rag-pipeline asset.
Test the connection
You can test the endpoint manually with curl before wiring up Langfuse:
curl -X POST "https://api.obsivara.com/ingest/webhook/langfuse?asset_id=my-langfuse-agent" \
-H "Authorization: Bearer obs_live_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{
"id": "trace_xyz789",
"name": "RAG pipeline",
"input": "What is the refund policy?",
"output": "You can request a refund within 30 days.",
"status": "success"
}'A successful response is HTTP 202:
{ "source": "langfuse", "accepted": 1, "rejected": 0, "errors": [] }Verify it worked
After triggering a Langfuse trace, open the Obsivara dashboard and find your asset. Traces typically appear within one minute. See Verify it worked for a full walkthrough.