Pinecone

Managed vector database for semantic search and retrieval-augmented generation. Use it to create indexes, generate embeddings and rerank results with Pinecone Inference, and upsert/query/fetch/delete vectors with metadata filtering. Unlike a general datastore, Pinecone is purpose-built for low-latency similarity search over high-dimensional embeddings at scale. 11 endpoints available through Lava’s AI Gateway. See the Pinecone API docs for full documentation.

This provider requires your own credentials — connect your API key or OAuth account before use.

Endpoints

List all indexes in the project

GET https://api.pinecone.io/indexes — Free

SDK
cURL

const data = await lava.gateway('https://api.pinecone.io/indexes', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.pinecone.io%2Findexes" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

Create a serverless index. `dimension` must match your embedding model (e.g. 1024 for multilingual-e5-large).

POST https://api.pinecone.io/indexes — Free

SDK
cURL

const data = await lava.gateway('https://api.pinecone.io/indexes', {
  body: {
"name": "my-index",
"dimension": 1024,
"metric": "cosine",
"spec": {
  "serverless": {
    "cloud": "aws",
    "region": "us-east-1"
  }
}
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.pinecone.io%2Findexes" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"name":"my-index","dimension":1024,"metric":"cosine","spec":{"serverless":{"cloud":"aws","region":"us-east-1"}}}'

Describe an index by name. Returns the `host` for data-plane requests (upsert/query/fetch/delete) on that index.

GET https://api.pinecone.io/indexes/my-index — Free

SDK
cURL

const data = await lava.gateway('https://api.pinecone.io/indexes/my-index', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.pinecone.io%2Findexes%2Fmy-index" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

Delete an index by name

DELETE https://api.pinecone.io/indexes/my-index — Free

SDK
cURL

const data = await lava.gateway('https://api.pinecone.io/indexes/my-index', { method: 'DELETE' });

curl -X DELETE "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.pinecone.io%2Findexes%2Fmy-index" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json"

Generate embeddings for input text with a hosted Pinecone model (e.g. multilingual-e5-large, llama-text-embed-v2)

POST https://api.pinecone.io/embed — Free

SDK
cURL

const data = await lava.gateway('https://api.pinecone.io/embed', {
  body: {
"model": "multilingual-e5-large",
"parameters": {
  "input_type": "passage",
  "truncate": "END"
},
"inputs": [
  {
    "text": "The quick brown fox"
  }
]
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.pinecone.io%2Fembed" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"multilingual-e5-large","parameters":{"input_type":"passage","truncate":"END"},"inputs":[{"text":"The quick brown fox"}]}'

Rerank documents against a query with a hosted reranking model (e.g. bge-reranker-v2-m3)

POST https://api.pinecone.io/rerank — Free

SDK
cURL

const data = await lava.gateway('https://api.pinecone.io/rerank', {
  body: {
"model": "bge-reranker-v2-m3",
"query": "what is a vector database",
"top_n": 2,
"documents": [
  {
    "id": "1",
    "text": "Pinecone is a vector database"
  },
  {
    "id": "2",
    "text": "A fox is an animal"
  }
]
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.pinecone.io%2Frerank" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"bge-reranker-v2-m3","query":"what is a vector database","top_n":2,"documents":[{"id":"1","text":"Pinecone is a vector database"},{"id":"2","text":"A fox is an animal"}]}'

Upsert vectors into a namespace. Target the index `host` from describe-index. `values` length must match the index dimension.

POST https://my-index-abc123.svc.aped-4627-b74a.pinecone.io/vectors/upsert — Free

SDK
cURL

const data = await lava.gateway('https://my-index-abc123.svc.aped-4627-b74a.pinecone.io/vectors/upsert', {
  body: {
"namespace": "ns1",
"vectors": [
  {
    "id": "v1",
    "values": [
      0.1,
      0.2,
      0.3
    ],
    "metadata": {
      "text": "hello"
    }
  }
]
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fmy-index-abc123.svc.aped-4627-b74a.pinecone.io%2Fvectors%2Fupsert" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"namespace":"ns1","vectors":[{"id":"v1","values":[0.1,0.2,0.3],"metadata":{"text":"hello"}}]}'

Query a namespace for the topK nearest vectors. Target the index `host` from describe-index. Supports metadata `filter`.

POST https://my-index-abc123.svc.aped-4627-b74a.pinecone.io/query — Free

SDK
cURL

const data = await lava.gateway('https://my-index-abc123.svc.aped-4627-b74a.pinecone.io/query', {
  body: {
"namespace": "ns1",
"vector": [
  0.1,
  0.2,
  0.3
],
"topK": 3,
"includeMetadata": true,
"includeValues": false
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fmy-index-abc123.svc.aped-4627-b74a.pinecone.io%2Fquery" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"namespace":"ns1","vector":[0.1,0.2,0.3],"topK":3,"includeMetadata":true,"includeValues":false}'

Fetch vectors by id from a namespace. Target the index `host` from describe-index. Pass `ids` and `namespace` as query params.

GET https://my-index-abc123.svc.aped-4627-b74a.pinecone.io/vectors/fetch?ids=v1&namespace=ns1 — Free

SDK
cURL

const data = await lava.gateway('https://my-index-abc123.svc.aped-4627-b74a.pinecone.io/vectors/fetch?ids=v1&namespace=ns1', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fmy-index-abc123.svc.aped-4627-b74a.pinecone.io%2Fvectors%2Ffetch%3Fids%3Dv1%26namespace%3Dns1" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

Delete vectors by id (or by metadata filter / deleteAll) from a namespace. Target the index `host` from describe-index.

POST https://my-index-abc123.svc.aped-4627-b74a.pinecone.io/vectors/delete — Free

SDK
cURL

const data = await lava.gateway('https://my-index-abc123.svc.aped-4627-b74a.pinecone.io/vectors/delete', { body: {"namespace":"ns1","ids":["v1","v2"]} });

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fmy-index-abc123.svc.aped-4627-b74a.pinecone.io%2Fvectors%2Fdelete" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"namespace":"ns1","ids":["v1","v2"]}'

Get vector counts and fullness per namespace for an index. Target the index `host` from describe-index.

POST https://my-index-abc123.svc.aped-4627-b74a.pinecone.io/describe_index_stats — Free

SDK
cURL

const data = await lava.gateway('https://my-index-abc123.svc.aped-4627-b74a.pinecone.io/describe_index_stats', { body: {} });

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fmy-index-abc123.svc.aped-4627-b74a.pinecone.io%2Fdescribe_index_stats" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

Get Started

AI Gateway

AI Spend

Monetize

Integration

SDK

Cookbook

Comparisons

Endpoints

List all indexes in the project

Create a serverless index. `dimension` must match your embedding model (e.g. 1024 for multilingual-e5-large).

Describe an index by name. Returns the `host` for data-plane requests (upsert/query/fetch/delete) on that index.

Delete an index by name

Generate embeddings for input text with a hosted Pinecone model (e.g. multilingual-e5-large, llama-text-embed-v2)

Rerank documents against a query with a hosted reranking model (e.g. bge-reranker-v2-m3)

Upsert vectors into a namespace. Target the index `host` from describe-index. `values` length must match the index dimension.

Query a namespace for the topK nearest vectors. Target the index `host` from describe-index. Supports metadata `filter`.

Fetch vectors by id from a namespace. Target the index `host` from describe-index. Pass `ids` and `namespace` as query params.

Delete vectors by id (or by metadata filter / deleteAll) from a namespace. Target the index `host` from describe-index.

Get vector counts and fullness per namespace for an index. Target the index `host` from describe-index.

Next Steps

All Providers

Forward Proxy

​Endpoints

​List all indexes in the project

​Create a serverless index. dimension must match your embedding model (e.g. 1024 for multilingual-e5-large).

​Describe an index by name. Returns the host for data-plane requests (upsert/query/fetch/delete) on that index.

​Delete an index by name

​Generate embeddings for input text with a hosted Pinecone model (e.g. multilingual-e5-large, llama-text-embed-v2)

​Rerank documents against a query with a hosted reranking model (e.g. bge-reranker-v2-m3)

​Upsert vectors into a namespace. Target the index host from describe-index. values length must match the index dimension.

​Query a namespace for the topK nearest vectors. Target the index host from describe-index. Supports metadata filter.

​Fetch vectors by id from a namespace. Target the index host from describe-index. Pass ids and namespace as query params.

​Delete vectors by id (or by metadata filter / deleteAll) from a namespace. Target the index host from describe-index.

​Get vector counts and fullness per namespace for an index. Target the index host from describe-index.

​Next Steps

All Providers

Forward Proxy

Endpoints

List all indexes in the project

Create a serverless index. `dimension` must match your embedding model (e.g. 1024 for multilingual-e5-large).

Describe an index by name. Returns the `host` for data-plane requests (upsert/query/fetch/delete) on that index.

Delete an index by name

Generate embeddings for input text with a hosted Pinecone model (e.g. multilingual-e5-large, llama-text-embed-v2)

Rerank documents against a query with a hosted reranking model (e.g. bge-reranker-v2-m3)

Upsert vectors into a namespace. Target the index `host` from describe-index. `values` length must match the index dimension.

Query a namespace for the topK nearest vectors. Target the index `host` from describe-index. Supports metadata `filter`.

Fetch vectors by id from a namespace. Target the index `host` from describe-index. Pass `ids` and `namespace` as query params.

Delete vectors by id (or by metadata filter / deleteAll) from a namespace. Target the index `host` from describe-index.

Get vector counts and fullness per namespace for an index. Target the index `host` from describe-index.

Next Steps