When you create a Pinecone Index, you choose the pod type and number of pods to use for the index. Each pod comes with a fixed amount of vCPU, RAM, and SSD coupled with the software. You will be billed at the end of the month for total pod-hours used. Pod usage is rounded up to 15-minute increments.
Pod-hours are counted for the time an index is running, starting from when the index is created. The number or duration of queries do not affect pod-hours.
Overview of pod types (see documentation for more details):
- s1: Optimized for storage and cost, with low overall cost and low throughput.
- p1: Optimized for balanced performance, with moderate search latency and throughput.
- p2: Optimized for high performance, with very low search latency and high throughput.
Google Cloud Platform
Amazon Web Services| Pod Type | Starter | Standard | Enterprise |
|---|---|---|---|
| s1 | Free (Limit: 1) | $0.0960 /pod-hour | $0.1440 /pod-hour |
| p1 | Free (Limit: 1) | $0.0960 /pod-hour | $0.1440 /pod-hour |
| p2 | Free (Limit: 1) | $0.1440 /pod-hour | $0.2160 /pod-hour |
| Pod Type | Starter | Standard | Enterprise |
|---|---|---|---|
| s1 | Free (Limit: 1) | $0.1110 /pod-hour | $0.1665 /pod-hour |
| p1 | Free (Limit: 1) | $0.1110 /pod-hour | $0.1665 /pod-hour |
| p2 | Free (Limit: 1) | $0.1665 /pod-hour | $0.24975 /pod-hour |
Scaling pod sizes to x2, x4, and x8 will change the pricing proportionally. For instance, a p1.x2 pod is twice the price of a p1 pod.
Contact us for pre-commitment discounts and Enterprise Dedicated pricing.
Common Questions
What will I be charged for?
You will be charged for “pod hours” in 15-minute increments, which is the time an index is running. Pod-hours are counted even when you’re not sending queries to the index. An index is “running” from the moment it’s created until the moment it’s deleted.
How many pods do I need?
Since each pod is a bundle of hardware resources (vCPU, RAM, and disk), the required number of pods is affected by multiple factors such as data size (volume, dimensionality, and metadata), metadata cardinality, and desired latency or throughput. We strongly recommend testing with your own data on different pod types and number of pods, and following our tips for performance tuning.
As a very rough starting guideline: each s1 pod fits 5M 768-dim vectors, each p1 pod fits 1M 768-dim vectors, and each p2 pod fits 1M 768-dim vectors.
You can also contact us for help with sizing and testing.
How secure is Pinecone?
Pinecone is SOC2 Type II compliant, and we take security seriously for all users on all plans. Read about our security practices.
Where is the user license or the terms of service?
Users accept the End-User License Agreement when creating an account.
Do you offer pre-commitment discounts?
Yes we do. Contact us for more details.
How do I cancel or downgrade?
You can manage your subscription directly in the console. If you attempt to downgrade to Starter (Free) while exceeding the limits of the Starter plan, you will first need to delete projects or indexes until you are within the limits.
For other questions, contact us or ask the community.