Pricing
Free to start. Then the environment, then the data you need
Signing up and getting an estimate costs nothing. After that: environment work from $800, charged once ever, then a connector per source system from $3,000 — including three data models. Every quote states its effort in 4-hour blocks, split into what is environment work and what is the connector.
First — sign up
Free
Free
Sign up, set up the basics and request a connector. We assess it and send you an estimate — what the environment work costs, what the connector costs, and how many hours each is. No card, no commitment.
- Your organisation, your team, your dashboard
- A code repository — the code is yours from the first commit
- Declare your warehouse and the source system you want connected
- Request a connector
- A written assessment: every dataset named, typed and explained
- An estimate in four-hour blocks, split into DevOps and Connector
The estimate is indicative until we have seen your API actually respond. We say so on the estimate.
Second — DevOps
Once per customer
Charged once, on your first engagement, and never again — your second connector has no DevOps line at all. Priced at $800 per 4-hour block, because what your environment needs depends on your environment.
Third — the connector
Per connector
$3,000 – $10,000+
Your API connected to your warehouse, with three Simple data models live. Authentication, scheduling, retries, rate limiting, state, monitoring and deployment — all handled, all yours.
- $3,0006 blocks
- Standard source. REST or GraphQL, documented, API key or OAuth2 client credentials, standard pagination, public network.
- $10,000+11 blocks
- Complex source. SOAP, session-token or mTLS authentication, private networking, undocumented pagination, or per-tenant credentials. Quoted individually above $10,000 where the source warrants it.
- The connector: auth, API access, scheduling, retries, rate limits, state, monitoring
- Three Simple data models of your choosing
- The raw API resources behind them, landed in your warehouse
- Source code, infrastructure definitions and documentation, in your accounts
One connector per source system. Every data model on it shares it, so a second data model never pays for the connector again.
Then add the data you need
Three Simple data models come with the Base Connector. Anything beyond that is priced by type — and the API resources behind each one are included, however many there are. The block count is our effort estimate, not a rate: the price is fixed whether or not it takes that long.
Simple
$1,500
1 block · 2–5 business days
3 included in the Base Connector
One resource, flat records, little or no transformation. Reference and lookup data, or a single entity with no dependencies.
e.g.Departments · Locations · Cost centres · Product categories
- One API resource
- Flat records
- Little or no transformation
- Three of these are included in a Base Connector
Standard
$5,000
3 blocks · 5–10 business days
A straightforward business entity. Normal reference data, simple relationships, standard incremental sync, limited transformation.
e.g.Employees · Customers · Products · Time entries
- One business entity
- Simple lookups and reference data
- Standard incremental synchronisation
- Limited transformation
Advanced
$12,500
9 blocks · 10–20 business days
Multiple dependent resources, complex relationships or nested structures, harder incremental sync, significant transformation.
e.g.Absences · Applicants · Orders with line items · Invoices
- Dependent resources, extracted in order
- Nested structures to flatten
- Harder incremental synchronisation
- Significant transformation
Custom
from$30,000
24 blocks · 20–40 business days
Substantial bespoke business logic, unusual reconciliation, more than one API, or exceptional technical requirements. Individually assessed.
e.g.Payroll · Reconciled financials · Cross-system entities
- Bespoke business logic
- Reconciliation across sources
- More than one API
- Exceptional technical requirements
Buying several, or adding support
Volume
- −10% 3–4 data models in one order
- −15% 5–9 data models in one order
- −20% 10 or more in one order
Support plan
- −20% with a 12-month support contract at the Standard tier or above with your order
pro-rata if support is cancelled before month 12. We would rather explain that clause up front than have it be a surprise.
The two stack. Both apply to additional data models only — never to DevOps or the connector, because those are already priced close to cost. Prices in USD.
Before you ask your security team
Where it runs, and who can reach it
The DevOps step is where a security review usually starts, so here are the five answers it usually needs. The short version: you choose where the connector runs, and you own it permanently either way.
- Where does the connector run?
- Your choice, and it is reversible. On our cloud (Hosted, Integrated), or on your own infrastructure including on-premise (Self-hosted). It is the same Docker image either way, so moving is a redeploy rather than a rebuild.
- Does our data pass through your systems?
- At Hosted and Integrated, yes — in transit. The connector extracts on our infrastructure and writes to your warehouse; nothing is retained after a successful run. At Self-hosted, no: nothing of yours reaches us at all.
- Are you a data processor?
- At the hosted levels, yes, and we will sign a DPA on that basis. At Self-hosted, no — we are not in the data path. We would rather you knew which of those you are buying before your review starts than after.
- Who holds our credentials?
- At Hosted, we do, in our secret manager. At Integrated and Self-hosted, you do — the connector resolves them at runtime from your secret manager or identity provider, and we hold nothing long-lived.
- What happens if we leave, or you disappear?
- You keep everything, and it keeps running. The connector is yours permanently, source in your repository. The SDK it is built against is yours too — the version you hold at the end is free to use, as-is, indefinitely. What ends is what was never in the container: our monitoring, our on-call, and updates as your vendor changes their API. Nothing switches off and there is no licence check in the image. That is true for scheduled and webhook connectors alike.
What you are buying
A connector, and the data models on it
A Base Connector is bought once per source system and comes with three data models. Everything after that is priced by type.
Connector
The technical connection
One source system into one warehouse — Personio → BigQuery. It handles authentication, API access, scheduling, retries, rate limiting, state management, monitoring and deployment.
Charged once per source system. Every data model on it shares it.
Data model
The data you actually asked for
A complete, usable business dataset in your warehouse — Employees, Absences, Applicants, Customers, Orders, Invoices. One data model may need several API resources; that is our problem, not your line item.
Three come with the Base Connector. After that they are priced by type — cheaper in bundles.
Worked example
One data model, four API resources, one price
And the raw resources land in your warehouse alongside the model.
What decides the type
The factors, in plain language
Assessed from the API's own documentation, with a written reason. If you disagree with a type, we show you the reason and re-run the classification — we do not split the price.
- How the resources relate
- One resource with simple lookups is Standard. Dependent resources that have to be extracted in a particular order push it to Advanced. The single most decisive question, and your own description usually answers it.
- Keeping it in sync
- A dataset the API can filter by change date is straightforward. One where we have to detect change ourselves, or where a child record changes without the parent saying so, is not.
- Nesting and shape
- Flat records cost nothing extra. Deeply nested structures that have to be flattened without losing meaning do.
- Transformation
- Typing and naming is included. Deriving business fields, resolving hierarchies or reconciling figures is what makes a dataset Advanced or Custom.
- History
- How much you need loaded. This is where rate limits, pagination depth and API inconsistency all surface at once.
- More than one API
- A dataset that reconciles data from two or more APIs is Custom, regardless of anything else.
- Grain
- Employees is one row per person; Absences is one row per event. Different grain means different data models — and it also means we will not split one dataset into two to charge twice.
- The connector underneath
- SOAP, session tokens, private networking or per-tenant credentials make the connector complex. That raises the connector line, once. It never pushes each data model up a type.
Not a price driver
What we never charge for
Stated plainly, because these are the five places a data pipeline quote usually grows.
- API endpoints
- An Employees data model reading /employees, /employee-fields, /departments and /locations is one data model, not four. Consumed resources are never separately priced.
- Warehouse tables
- A data model may produce several tables. How many is a modelling consequence, never a price.
- Data volume
- 500 rows a day and 5 million rows a day are usually the same data model. Volume changes your cloud bill, which you pay directly, and occasionally the architecture. It does not change our price by itself.
- Raw data delivery
- The raw resources behind a data model are always included. They are not extra data models and they are not a line item.
- The connector, more than once
- A Base Connector is bought once per source system. Ten data models on one source pay for the connector once — not ten times.
Optional support
Separate, and never required
The connector runs whether or not you buy this. If you ever feel you have to pay us for it to keep working, we have mis-sold it.
| What you pay for | Unit | Price |
|---|---|---|
| Account | per month | $200 |
| Operations — monitoring, logs, alerting, the infrastructure behind them | per connector, per month | from $100 |
| Guaranteed hours — you decide what they are spent on | per hour, per month | $175 |
| Plan | Guaranteed hours / month | Over the 12 months | Per month, one connector |
|---|---|---|---|
| Essential | 2 | 24 | $650 |
| Standard | 4 | 48 | $1,000 |
| Extended | 8 | 96 | $1,700 |
| Programme | 16 | 192 | $3,100 |
Every plan answers inside 2 business days, in European business hours, 08:00–18:00 CET, Mon–Fri — they differ only in hours. Faster response and wider coverage are adjustments we price on request, not plans you have to move up to. The minimum contract is 2 hours a month over 12 months — 24 hours, $650/month — cancellable with 30 days' notice after month 3. Unused hours carry one month forward and then expire. There is no 24/7 tier — not gated behind a threshold, simply not offered.
Your hours are for
- Connector updates and minor changes
- Absorbing a vendor schema change within the same resources
- Investigating and fixing breakage
- Tuning: cadence, batching, backfill windows, rate-limit behaviour
- Additional monitors, alerts and runbook entries
- Re-running a backfill
Included, and never taken out of your hours
- We receive and triage the alerts
- Run-health reporting
- Security patching
- Dependency and library updates
- SDK version upkeep
- The two-business-day response target itself
Keeping your connector patched and its libraries current is our job, not something you should be spending your hours on. Nor is fixing a defect in what we built.
Not included
What the fixed price does not cover
Stated here rather than discovered later.
- Your cloud costs
- Paid by you to your cloud provider. Typically $5–25/month for a daily sync, $25–90 hourly. We take no margin on infrastructure.
- The analytics layer
- We deliver Employees. Your headcount dashboard, your metric definitions and your dbt marts on top of it are yours.
- Building the warehouse
- You need BigQuery, Snowflake or Databricks already. We are the wrong first call otherwise.
- A new resource in an existing data model
- Quoted, typically 2 blocks — $1,600. A genuinely new business dataset is a new data model at its assessed type, not a change request.
- Breaking changes by the API vendor
- New work, quoted. A support contract's guaranteed hours absorb schema changes within the same resources; a breaking change is quoted in every case.
Minimum engagement $3,000 — one Standard data model. Below that, the cost of quoting and contracting exceeds what the work earns, and we would rather say so.
Get an estimate
Name the data. Send the docs.
Tell us which datasets you need and link the API documentation. We work out the resources behind each one, classify it, and send back a fixed price, a delivery window and your expected monthly cloud cost. Free, usually the same day.
Not sure it's a custom job?
Send it anyway. If Fivetran, Airbyte or the vendor already has a connector that gives you the data you need, we will tell you and you should use it — that answer is free too.
Ask us