Transparent and Predictable Pricing for the long haul
We believe in being transparent, simple and predictable.
Free Tier
Perfect for getting started and small projects
Pay as You Go
Transparent all-in-one pricing
How we calculate ingestion
Eval loop and data ingestion
We measure the data going into our eval loop, which serves as a proxy for various usage metrics.
Every effect within a workflow results from an eval loop, triggered by actions or model reasoning. We measure the input size to the nearest byte on each eval iteration.
A byte typically corresponds to a single character, but may vary based on encoding.
Pricing Example
Let's say you have a workflow that starts with a webhook trigger and 100 bytes of data. To arrive at a conclusion, the model reasons through 5 iterations, adding 100 bytes to the workflow context each time.
- First iteration: 100 bytes
- Second iteration: 200 bytes
- Third iteration: 300 bytes
- Fourth iteration: 400 bytes
- Fifth iteration: 500 bytes
- Total ingestion: 1,500 bytes (~1.5 KB)
If you have 1000 of these workflows running per day, you'll be billed 1.5 KB * 1000 = 1500 KB per day.
At $0.006 per KB, that's $9 per day, $270 per month.
Our Approach to Pricing
Transparency and simplicity
We believe in keeping our pricing model straightforward and easy to understand. We measure the data going into our eval loop, which acts as a proxy for other usage metrics like compute, tokens, egress and storage.
Incentivizing best product experience
Our pricing is based on a simple, flat rate for the amount of data we ingest. This approach gives you direct control over your costs by allowing you to manage your functions and workflow behavior.
Our pricing incentivizes a configuration that provides the best product experience. A large number of shorter workflows are significantly cheaper than a smaller number of longer workflows. They also lead to better accuracy, reliability, and repeatability of results.
Sustainability in operation
We don't price based on tokens because we use a mix of models under the hood, employing complex model routing to achieve the best results. Our goal is to provide value and make a fair profit from each paying customer, without relying on loss leaders or additional professional services.