Neatlogs
Quickstart

Migrate to Neatlogs

Import your traces from Langfuse, LangSmith, Braintrust, or Raindrop.

Bring your existing traces into Neatlogs from four observability platforms: Langfuse, LangSmith, Braintrust, and Raindrop. Set up one-time imports or recurring syncs to keep your data in sync.

Setting Up an Import

All imports start from Settings → Integrations.

Step 1: Choose Your Source

  1. Go to SettingsIntegrations.
  2. Select your source platform:
    • Langfuse - Production-ready LLM observability
    • LangSmith - LangChain's tracing platform
    • Braintrust - Evals + observability
    • Raindrop - Traceloop's trace aggregation

Step 2: Create a Connection

Choose your source platform and gather the required API credentials:

CredentialDetails
Public KeyFind in your Langfuse project settings
Secret KeySame location as Public Key
Base URLOptional — defaults to https://cloud.langfuse.com
CredentialDetails
API KeyGet from LangSmith settings
Base URLOptional — defaults to https://api.smith.langchain.com
CredentialDetails
API KeyGet from Braintrust settings
Base URLOptional — defaults to https://api.braintrust.dev
CredentialDetails
API KeyGet from Raindrop Query API — 96-character hex key

Step 3: Select Projects to Import

After authenticating, Neatlogs discovers your source projects. Choose which ones to import.

Step 4: Configure Import Settings

  • Import mode: One-time or recurring
    • One-time - Import traces up to a specific date
    • Recurring - Keep syncing on a schedule (daily, hourly, custom)
  • Date range: How far back to import (default: last 7 days)

Monitoring Import Progress

Once an import job starts, you can track it in Integrations.

Each import shows:

  • Status - pending, running, succeeded, failed, paused
  • Progress - traces processed vs. total discovered
  • Stats - traces imported, replaced, skipped, failed
  • Last sync - when the import last ran
  • Error messages - if something went wrong

Recurring Imports

Recurring imports run automatically on your configured schedule. They:

  • Only sync new/updated traces since the last run (delta sync)
  • Skip traces that haven't changed
  • Never duplicate data
  • Can be paused or cancelled anytime

Imported Traces in Neatlogs

Imported traces are marked with a source badge showing where they came from (Langfuse, LangSmith, etc.). They appear in:

  • Traces list - Filterable by import source
  • AI Search - Query across imported and native traces
  • Analytics - Track metrics on imported data
  • Comments - Collaborate on imported traces like native ones

You can filter by source in the traces list, or use the tag imported:<source> in search queries:

imported:langfuse error_rate > 0.1
imported:braintrust cost > 10

Troubleshooting Imports

Your API credentials may be invalid or expired.

How to fix:

  1. Verify your API credentials are correct and match your source settings
  2. Check that your source account still has active access
  3. Ensure your API key hasn't been revoked or rotated
  4. Test the connection in Integrations before running the import

The import completed but you don't see traces.

How to fix:

  1. Check that the date range captures when your traces were created
  2. Verify you selected the correct source project (not just the account)
  3. Wait for the import job to complete — check status in Integrations
  4. Check for error messages in the import job details
  5. Try importing a smaller date range first to isolate the issue

The import job failed or was paused unexpectedly.

How to fix:

  1. Open Integrations and check the error message for details
  2. Verify your API key is still valid (hasn't expired)
  3. Check your source platform's status (may be down)
  4. Resume or restart the import from the Integrations page
  5. If it keeps failing, contact support with the error message

You see duplicate traces after importing.

How to fix: Neatlogs automatically deduplicates traces. Duplicates you see may be:

  • Different source projects with the same trace IDs
  • Traces with identical content but different timestamps
  • Intentional re-imports from an earlier date range

If you suspect a deduplication bug, contact support with trace IDs that appear duplicated.

On this page