What authentication method is used?

HTTP Basic Auth with the public key as the username and the secret key as the password.

Which endpoints can I query?

Endpoints include health, traces, and observations, with detail endpoints for traces and observations.

How do I filter results by time or model?

Use fromTimestamp and toTimestamp to filter traces by time, and filter generations by the model field in the response.

langfuse-observability

npx machina-cli add skill ddnetters/homelab-agent-skills/langfuse-observability --openclaw

Files (1)

SKILL.md

7.8 KB

Langfuse Observability

Langfuse is an open-source LLM observability platform. Use this skill to query traces, generations, costs, and metrics from a self-hosted or cloud Langfuse instance.

Setup

# Set your Langfuse credentials
export LANGFUSE_HOST="https://langfuse.example.com"  # or https://cloud.langfuse.com
export LANGFUSE_PUBLIC_KEY="pk-lf-..."
export LANGFUSE_SECRET_KEY="sk-lf-..."

All API calls use basic auth with public key as username and secret key as password.

# Base pattern for all requests
curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/<endpoint>"

Health check

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/health" | jq .

Traces

List recent traces

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/traces?limit=10&page=1" | \
  jq '.data[] | {id, name, timestamp, userId, totalCost, latency, tags}'

Filter by time range

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/traces?limit=20&fromTimestamp=2026-02-01T00:00:00Z&toTimestamp=2026-02-28T23:59:59Z" | \
  jq '.data[] | {id, name, timestamp, totalCost, latency}'

Filter by user or name

# By user
curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/traces?limit=10&userId=USER_ID" | \
  jq '.data[] | {id, name, timestamp, totalCost}'

# By trace name
curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/traces?limit=10&name=TRACE_NAME" | \
  jq '.data[] | {id, name, timestamp, totalCost}'

Get trace details

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/traces/TRACE_ID" | \
  jq '{id, name, timestamp, userId, sessionId, totalCost, latency, tags, metadata,
       observations: [.observations[] | {id, type, name, model, startTime, totalCost, usage}]}'

Get trace input/output

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/traces/TRACE_ID" | jq '{input, output}'

Generations (LLM calls)

List recent generations

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/observations?type=GENERATION&limit=10&page=1" | \
  jq '.data[] | {id, name, model, startTime, totalCost, inputUsage: .usage.input, outputUsage: .usage.output}'

Filter by model

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/observations?type=GENERATION&limit=10" | \
  jq '[.data[] | select(.model | test("gpt-4"))] | .[] | {id, model, totalCost}'

Get generations for a trace

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/observations?type=GENERATION&traceId=TRACE_ID" | \
  jq '.data[] | {id, name, model, totalCost, usage, input, output}'

Generation details

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/observations/OBSERVATION_ID" | \
  jq '{id, type, name, model, modelParameters, startTime, endTime, totalCost, usage, level, statusMessage}'

Errors and warnings

# Find errors
curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/observations?level=ERROR&limit=10" | \
  jq '.data[] | {id, traceId, name, model, level, statusMessage, startTime}'

# Find warnings
curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/observations?level=WARNING&limit=10" | \
  jq '.data[] | {id, traceId, name, model, level, statusMessage, startTime}'

Daily metrics

Cost and usage by day

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/metrics/daily?limit=7" | \
  jq '.data[] | {date, traces: .countTraces, cost: .totalCost,
      models: [.usage[] | {model, input: .inputUsage, output: .outputUsage, total: .totalUsage}]}'

Total cost summary

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/metrics/daily?limit=30" | \
  jq '{totalCost: ([.data[].totalCost] | add), totalTraces: ([.data[].countTraces] | add),
       days: (.data | length), breakdown: [.data[] | {date, cost: .totalCost, traces: .countTraces}]}'

Sessions

# List recent sessions
curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/sessions?page=1&limit=10" | \
  jq '.data[] | {id, createdAt, projectId}'

# Get session details
curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/sessions/SESSION_ID" | jq .

Scores

# List scores
curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/scores?limit=10" | \
  jq '.data[] | {id, name, value, dataType, traceId, comment}'

# Scores for a specific trace
curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/scores?traceId=TRACE_ID" | \
  jq '.data[] | {name, value, dataType, comment}'

Docker self-hosted deployment

services:
  langfuse-web:
    image: langfuse/langfuse:3
    ports:
      - "3000:3000"
    environment:
      DATABASE_URL: postgresql://user:pass@postgres:5432/langfuse
      NEXTAUTH_SECRET: your-secret
      SALT: your-salt
      ENCRYPTION_KEY: your-encryption-key
      NEXTAUTH_URL: https://langfuse.example.com
      AUTH_DISABLE_SIGNUP: "true"
    depends_on: [postgres]

  langfuse-worker:
    image: langfuse/langfuse-worker:3
    environment:
      DATABASE_URL: postgresql://user:pass@postgres:5432/langfuse
      SALT: your-salt
      ENCRYPTION_KEY: your-encryption-key

  postgres:
    image: postgres:17
    volumes:
      - langfuse_pg:/var/lib/postgresql/data
    environment:
      POSTGRES_DB: langfuse
      POSTGRES_USER: user
      POSTGRES_PASSWORD: pass

Langfuse v3 also supports ClickHouse for trace storage, Redis for queuing, and MinIO for object storage.

OpenRouter integration

Send traces automatically from OpenRouter to Langfuse:

Go to https://openrouter.ai/settings/integrations
Enable Langfuse destination
Set Base URL: https://langfuse.example.com
Add Public Key and Secret Key from Langfuse Settings → API Keys

All OpenRouter LLM calls will then appear as traces in Langfuse.

Common queries

Most expensive traces this week

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/traces?limit=50&fromTimestamp=$(date -d '7 days ago' +%FT00:00:00Z)" | \
  jq '[.data[] | {id, name, totalCost, timestamp}] | sort_by(-.totalCost) | .[:10]'

Model usage breakdown

curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/observations?type=GENERATION&limit=100" | \
  jq '[.data[] | .model] | group_by(.) | map({model: .[0], count: length}) | sort_by(-.count)'

Error rate

ERRORS=$(curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/observations?level=ERROR&limit=1" | jq '.meta.totalItems')
TOTAL=$(curl -s -u "$LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY" \
  "$LANGFUSE_HOST/api/public/observations?limit=1" | jq '.meta.totalItems')
echo "Error rate: $ERRORS / $TOTAL"

Tips

Always use time filters (fromTimestamp/toTimestamp) on large queries to avoid timeouts
Pagination: responses include meta.page, meta.limit, meta.totalItems, meta.totalPages
Timestamps use ISO 8601 format: 2026-02-10T00:00:00Z
Cost values are in USD
Token counts are in usage.input, usage.output, usage.total
The web UI at your Langfuse URL provides visual exploration of traces

Source

git clone https://github.com/ddnetters/homelab-agent-skills/blob/main/langfuse-observability/SKILL.mdView on GitHub

Overview

Langfuse Observability lets you query traces, generations, costs, and metrics from a self-hosted or cloud Langfuse instance via REST API. It supports health checks and deep inspection of LLM pipelines, including per-trace usage and total costs.

How This Skill Works

Authenticate with HTTP Basic Auth using LANGFUSE_HOST and your public/secret keys, then call endpoints under /api/public (for traces, health, observations). The API returns JSON you can parse to extract fields like totalCost, latency, and usage, including nested observations for each trace.

When to Use It

Audit recent traces for latency, totalCost, and tags
Filter traces by time range or by user/name to locate activity
Inspect a trace's input/output and metadata for debugging
Review generations and their costs per model
Run a health check and surface errors or warnings in the Langfuse instance

Quick Start

Step 1: Set up credentials by exporting LANGFUSE_HOST, LANGFUSE_PUBLIC_KEY, and LANGFUSE_SECRET_KEY
Step 2: Use the base curl pattern: curl -s -u $LANGFUSE_PUBLIC_KEY:$LANGFUSE_SECRET_KEY $LANGFUSE_HOST/api/public/<endpoint>
Step 3: Example: list recent traces with the traces endpoint, e.g. $LANGFUSE_HOST/api/public/traces?limit=5 and inspect fields

Best Practices

Store LANGFUSE_HOST and keys in environment variables and rotate keys regularly
Always use HTTPS and curl with -u for secure authentication
Paginate results with limit and page when querying large datasets
Filter by fromTimestamp/toTimestamp to narrow the results to a period
Use jq or your language tooling to extract fields like totalCost, latency, and model

Example Use Cases

List the 10 most recent traces and review id, name, timestamp, totalCost, and latency
Get detailed data for a specific trace, including its input, output, and observations
Fetch generations for a trace and summarize totalCost and per-model usage
Filter generations by model (for example gpt-4) to compare costs
Run a health check against your Langfuse instance and surface status and errors

Frequently Asked Questions

Add this skill to your agents