PostHog Sources datasheet
Contents
What is PostHog Sources?
PostHog Sources is a fully managed tool that extracts data from your apps and databases and loads it into any data warehouse. It isn't reliant on PostHog at all: your data can go from Postgres to Snowflake, or Stripe to BigQuery, without ever touching PostHog. We handle pipeline maintenance like schema changes, incremental syncs, retries and rate limits.
Key messages
- Move data anywhere. PostHog Sources syncs Stripe, Postgres, HubSpot, Salesforce and the rest of your stack into any data warehouse, including PostHog's.
- Stop maintaining pipelines. Pipeline maintenance is 70% of a data engineer's job. We handle schema changes, incremental syncs, retries and rate limits. No more being on-call for a broken connector or wondering if your data is accurate and up to date.
- Predictable pricing. Per-row pricing is published on our pricing page; it gets cheaper as you grow and has no per-seat charges. No surprise invoices, and a fraction of the cost of competitors who have opaque pricing that results in surprise bills for users.
Key features
Move data in minutes, set and forget.
- 700+ sources: Stripe, HubSpot, Salesforce, Zendesk, Postgres and more, with new connectors shipping regularly.
- AI-assisted setup: Paste in a read-only key and pick your tables, or run
npx @posthog/wizard warehouseand the wizard finds your databases and APIs for you. - Free historical syncs: Every row synced in a new source's first 7 days is free, so backfills don't land on the bill.
- A reliable sync engine: Every incremental sync runs on deltalite, our own upsert engine. Memory scales with the batch, not the table, so big tables sync reliably.
- Secure by default: Credentials are encrypted at rest, you can connect with a read-only user, and PostHog is SOC 2 certified with EU data residency and a HIPAA BAA available.
Sources, or ELT?
Sources is our name for it. The rest of the industry calls this kind of tool ELT (or ETL). Data people search for, compare and budget for "ELT tools". We call it Sources because we aren't solely serving data professionals, when were targeting product engineers we should talk primarily about Sources as ELT would be an unknown and confusing term.
ELT stands for Extract, Load, Transform:
- Extract: Pull data out of where it lives, like Stripe, Postgres or HubSpot.
- Load: Copy it, as it is, into a warehouse.
- Transform: Clean, join and model it inside the warehouse with SQL, once it's there. This is not technically covered by our tool but by whatever warehouse the user loads their data into.
Older ETL tools transformed data before loading it. ELT loads the raw data first and transforms it later, so you keep everything and can change your mind.
How does it work?
Connect a source with a read-only key, pick the tables you want and the destination you want to send it to, and the first sync starts straight away. After that, PostHog keeps the copy up to date on a schedule that you set and fixes things when an API changes or has a bad day.
Common use cases
Apps to warehouse
| Use case | Who it's for | What changes |
|---|---|---|
| Replace an expensive tool like Fivetran or Airbyte | Data leads paying Fivetran or running Airbyte | Very similar usage experience. At 100M rows a month the saving is roughly $1,500 to $2,150. |
| Get data into a warehouse to model and query | Data teams pulling apps and databases into one place | Every source lands in one warehouse as raw tables, ready to model with SQL views and query together, with no pipelines to build or maintain. |
Data to PostHog
| Use case | Who it's for | What changes |
|---|---|---|
| Revenue and churn analysis | Teams with Stripe or billing data | Stripe syncs next to product events, so you can see how usage turns into revenue. |
| Sales and CRM reporting | Teams on HubSpot or Salesforce | Track the sales funnel with SQL alongside product usage, with no export scripts to maintain. |
| Marketing analytics | Marketing teams running paid ads | Ad spend and campaign data sync into PostHog next to product events for marketing analytics. |
Buying signals
Established data stack: ready to replace Fivetran or Airbyte
They already have an external warehouse and the tools around it. Pitch PostHog Sources as a cheaper, fully managed alternative to the ELT tool they use today.
- They already use batch exports: They send PostHog data out to their own warehouse.
Early days: ready to connect their first source
They use PostHog but have no warehouse or other data tools yet. Pitch PostHog Sources as the quickest way to get business data next to their product data.
- They use PostHog but have no sources connected: Everything they analyze today comes from events alone.
- They ask questions events can't answer: Revenue, churn, plan tier or which ad campaigns bring in good users.
- Their event volume is growing: This indicates their business has customers, which equals revenue. Time to connect Stripe to PostHog!
- They join data by hand: They repeatedly export CSVs from PostHog.
PostHog vs Fivetran vs Airbyte
| PostHog Sources | Fivetran | Airbyte Cloud | |
|---|---|---|---|
| 100M rows/month (list) | $735 | $2,894 (Standard, one Postgres connection) | $2,226 (Plus, 603 credits); $3,010 on Standard |
| Free tier | 1M rows/month, every month | 500,000 MAR | None; 30-day trial with 400 credits |
| Pricing unit | Rows synced | Monthly active rows, per connection | Credits (rows for APIs, GB for databases and files) |
| Minimum spend | None | $5/month per connection | $20/month on Standard |
| Historical backfills | Free for each new source's first 7 days | Count towards usage | Count towards credits |
| Managed schema changes, retries, rate limits | Yes | Yes | Yes |
| Data usable in analytics, experiments and agents | Yes, out of the box | No, needs a separate BI and product stack | No, needs a separate BI and product stack |
Where they're stronger
- Both have a better brand recognition and well established footholds. Point to PostHogs established rputation in other fields and entering new markets.
- Airbyte's open-source version is free if you host it, but someone has to run and fix it.
Pricing framework
- The first 1 million rows synced every month are free.
- Historical syncs are free: every row synced in a new source's first 7 days costs nothing. See the pricing calculator.
| Monthly rows synced | Price per row | Monthly cost at the top of the tier (list) |
|---|---|---|
| First 1M | Free | $0 at 1M |
| 1M–10M | $0.000015 | $135 at 10M |
| 10M–25M | $0.00001 | $285 at 25M |
| 25M–50M | $0.000008 | $485 at 50M |
| 50M–100M | $0.000005 | $735 at 100M |
| 100M–1B | $0.000002 | $1,535 at 500M; $2,535 at 1B |
| Above 1B | $0.000001 | $2,535 + $1 per extra 1M rows |
Competitor pricing
- Fivetran bills monthly active rows (MAR): distinct rows inserted, updated or deleted, counted per connection. Since January 2026, deletes count towards MAR and every connection carries a $5 monthly minimum, which raised many customers' bills. Each source costs a different amount per MAR.
- Airbyte bills credits, not rows. A credit buys a different amount of work depending on the source (rows for APIs, GB for databases and files). Standard is $20 a month plus $5 per extra credit; Plus is sold in credit packages.
Quick links and who to ask
#team-warehouse-pipelinesfor questions and support@lizzieeptonfor marketing and sales resource requests- Sources docs
- Migration guide:
- Wizard setup command:
npx @posthog/wizard warehouse - Connect your first source in app