alexai-mcp/hermes-ccc/pinecone

pip install pinecone-client sentence-transformers openai

from pinecone import Pinecone, ServerlessSpec
pc = Pinecone(api_key="your-api-key")  # or os.environ["PINECONE_API_KEY"]

# Serverless (pay-per-query — cheapest to start)
pc.create_index(
    name="my-index",
    dimension=1536,          # match your embedding model
    metric="cosine",         # cosine | euclidean | dotproduct
    spec=ServerlessSpec(
        cloud="aws",
        region="us-east-1"
    )
)
# Connect to index
index = pc.Index("my-index")

from sentence_transformers import SentenceTransformer
model = SentenceTransformer("all-MiniLM-L6-v2")  # dim=384
documents = [
    {"id": "doc1", "text": "Python async programming guide"},
    {"id": "doc2", "text": "Machine learning with PyTorch"},
]
vectors = []
for doc in documents:
    embedding = model.encode(doc["text"]).tolist()
    vectors.append({
        "id": doc["id"],
        "values": embedding,
        "metadata": {"text": doc["text"], "source": "manual"}
    })
# Batch upsert (max 100 per call)
index.upsert(vectors=vectors, namespace="docs")

query_text = "how to write async Python?"
query_vector = model.encode(query_text).tolist()
results = index.query(
    vector=query_vector,
    top_k=5,
    namespace="docs",
    include_metadata=True,
)
for match in results["matches"]:
    print(f"Score: {match['score']:.3f} | {match['metadata']['text']}")

results = index.query(
    vector=query_vector,
    top_k=5,
    filter={"source": {"$eq": "manual"}},
    include_metadata=True,
)
# Operators: $eq, $ne, $gt, $gte, $lt, $lte, $in, $nin, $and, $or
results = index.query(
    vector=query_vector,
    top_k=5,
    filter={
        "$and": [
            {"category": {"$in": ["tech", "science"]}},
            {"year": {"$gte": 2023}},
        ]
    },
    include_metadata=True,
)

# Different namespaces = separate vector spaces (free, no extra cost)
index.upsert(vectors=vectors, namespace="user-123")
index.upsert(vectors=vectors, namespace="user-456")
# Query specific namespace
results = index.query(vector=query_vector, top_k=5, namespace="user-123")
# Delete namespace
index.delete(delete_all=True, namespace="user-123")

# Fetch specific vectors
fetched = index.fetch(ids=["doc1", "doc2"], namespace="docs")
# Delete vectors
index.delete(ids=["doc1"], namespace="docs")
# Update metadata (re-upsert with same id)
index.upsert(vectors=[{"id": "doc1", "values": embedding, "metadata": {"updated": True}}])

stats = index.describe_index_stats()
print(f"Total vectors: {stats['total_vector_count']}")
print(f"Namespaces: {stats['namespaces']}")