Skip to content
MyStackPlan
Vector Databases

Pinecone pricing

A managed vector database for semantic search and AI apps. Has a free plan, a $20 flat plan and a pay-as-you-go plan with a $50 monthly minimum.

Try it in the builderOfficial pricing pageReport a wrong price

Every plan below was checked against Pinecone's own pages on . Prices in USD.

Plans

Starter

Free

At the limit
Stops at the limit
Limits
  • Vector storage: 2 GB
  • Vector queries: 4M queries a month1M read units a month, at the 0.25 read-unit minimum per query
  • Free limits: 2 GB storage, 1M read units, 2M write units and 1 GB egress a month.
  • Past a monthly limit requests fail with a 429 error until you upgrade. Past the storage limit you can delete records instead.
  • Starter indexes must run on AWS us-east-1. Limited to 1 project, up to 5 indexes and 2 users.
  • A query uses 1 read unit per GB of namespace size, with a minimum of 0.25. Splitting data into one namespace per user keeps queries at the minimum; one big namespace costs more.
  • Writes use 1 write unit per KB. They are not in this estimate.

Source: pinecone.io · checked

Builder

$20.00/month

At the limit
Stops at the limit
Limits
  • Vector storage: 10 GB
  • Vector queries: 8M queries a month2M read units a month, at the 0.25 read-unit minimum per query
  • A flat $20 a month covers 10 GB storage, 2M read units, 5M write units and 10 GB egress.
  • Usage past these limits is blocked rather than billed, so you cannot get a surprise bill. You must move to Standard to grow.
  • Includes up to 5 projects, 5 users, any region on AWS, Azure or GCP, and 10 indexes per project.

Source: pinecone.io · checked

Standard

$50.00/month

Includes $50.00 of usage credit a month

At the limit
Bills for extra usage
Usage pricing
  • Vector storage: $0.33 per GB
  • Vector queries: $4 per 1M queries
  • Minimum trap: you are charged at least $50 a month. Usage below $50 is covered by that minimum, and you pay for actual usage only once it goes over.
  • Read units cost $16 to $18 per million and write units $4 to $4.50 per million, depending on cloud and region. This estimate uses the lowest rate.
  • Writes (1 write unit per KB, so about 6 write units per 6 KB vector) and egress above 100 GB ($0.10 per GB) are not in this estimate. Loading many vectors can cost far more than storing them.
  • A query uses 1 read unit per GB of namespace size, with a minimum of 0.25. This estimate assumes small per-user namespaces.
  • A 3-week trial includes $300 of credits.

Source: pinecone.io · checked

Works well with

  • VercelPinecone has a Vercel integration that connects Vercel AI projects to a Pinecone index.

Alternatives for vector databases