Linking Helicone as a source
Let AI connect your sources for you
Skip the manual setup — run this in your project and the wizard auto-detects your databases and APIs and connects them to PostHog.
Contents
The Helicone connector syncs your logged LLM requests, sessions, per-user usage aggregates, and prompts from Helicone into PostHog. This lets you join LLM cost, token usage, and latency data with your product analytics.
Prerequisites
You need a Helicone account with an API key. Any Helicone plan works, but the amount of request history you can sync depends on your plan's data retention.
Adding a data source
- In PostHog, go to the Sources tab of the data pipeline section.
- Click + New source and click Link next to this source.
- Enter your credentials (see Configuration below) and click Next.
- Select the tables you want to sync, choose a sync method and frequency, then click Import.
Once the syncs are complete, you can start querying this data in PostHog.
To connect, you need:
- API key – generate one in your Helicone dashboard under Settings → API Keys.
- Region – pick US or EU to match where your Helicone organization is hosted.
Sync modes
Each table can be synced in one of several modes, depending on what the source supports:
- Webhook (when available) – the source pushes changes to PostHog in real time. Fastest freshness, lowest ongoing cost, and the only mode that reliably captures updates and deletes.
- Incremental – only new or updated rows are synced on each run, using a cursor field (such as an
updated_attimestamp). Cheaper than a full refresh, but deletes aren't captured. - Append only – new rows are appended using a cursor field; existing rows are never updated. Ideal for immutable, append-only tables like event logs.
- Full refresh – the whole table is reloaded on every sync. Use it when a table has no reliable cursor or when you need deletions reflected.
See sync methods for a full explanation of how each mode works and how to choose between them.
The requests table is an append-only log, so incremental sync on request_created_at is recommended. Only the last 365 days are pulled on the initial incremental sync; subsequent runs fetch only newer requests. The sessions, users, and prompts tables contain aggregates and metadata that change in place, so they sync as full refreshes.
Configuration
| Option | Type | Required |
|---|---|---|
API key | password | Yes |
Region | select | Yes |
Supported tables
| Table | Description | Sync method | Incremental field | Primary key |
|---|---|---|---|---|
requests | Only syncs the last 365 days on initial incremental sync | Incremental, Full refresh | request_created_at | — |
sessions | Aggregated metrics for each Helicone session (requests grouped via the Helicone-Session-Id header): totals for cost, requests, and tokens, plus the users involved. | Full refresh | — | — |
users | Per-user usage aggregates across all logged requests: request count, token usage, and cost, keyed by the Helicone-User-Id header value. | Full refresh | — | — |
prompts | Prompts managed in Helicone's prompt registry, with their names and tags. | Full refresh | — | — |
properties | Catalog of the custom property keys (set via Helicone-Property-* headers) seen on the organization's requests, excluding keys hidden in Helicone. | Full refresh | — | — |
eval_scores | Catalog of the evaluation score names recorded on the organization's requests. | Full refresh | — | — |
Sync limitations
Request and response bodies above Helicone's size limits are offloaded to storage and referenced by a short-lived signed_body_url on each request row, rather than being embedded in the synced data. The per-user users table is an aggregate over your full request history and is fetched in a single request, without pagination.
Troubleshooting
If credential validation fails, check that the key was created in the same Helicone region you selected – a US key won't authenticate against the EU host, and vice versa.
If your sync is failing or data looks wrong, see the Data warehouse troubleshooting guide. If that doesn't help, contact support – we're happy to help.