Python is the most popular language for data analysis, trading bots, and automation. This tutorial shows you how to connect to Yellowstone gRPC from Python, subscribe to real-time Solana data, and handle events — or skip the complexity entirely with Subglow's pre-parsed JSON output.

Why Python for Solana Streaming

Python's strength isn't raw speed — it's the ecosystem. Pandas for data analysis, NumPy for computation, scikit-learn and TensorFlow for ML-based trading strategies. The bottleneck for Solana bots isn't Python's processing speed; it's data delivery latency. Yellowstone gRPC delivers data in ~5ms. Even in Python, processing each event takes microseconds.

Install Dependencies

pip install grpcio grpcio-tools protobuf

Clone the Yellowstone proto files:

git clone https://github.com/rpcpool/yellowstone-grpc.git
cd yellowstone-grpc/yellowstone-grpc-proto/proto

Generate Python stubs:

python -m grpc_tools.protoc \
  --proto_path=. \
  --python_out=./generated \
  --grpc_python_out=./generated \
  geyser.proto

This generates geyser_pb2.py and geyser_pb2_grpc.py — the Python bindings for Yellowstone's protobuf API.

Connect and Subscribe

import grpc
from generated import geyser_pb2, geyser_pb2_grpc

# Create secure channel
channel = grpc.secure_channel(
    "grpc.subglow.io:443",
    grpc.ssl_channel_credentials(),
)
stub = geyser_pb2_grpc.GeyserStub(channel)

# Build subscription request
def build_request():
    request = geyser_pb2.SubscribeRequest()
    # Subscribe to Pump.fun transactions
    tx_filter = geyser_pb2.SubscribeRequestFilterTransactions()
    tx_filter.account_include.append("6EF8rrecthR5Dkzon8Nwu78hRvfCKubJ14M5uBEwF6P")
    tx_filter.vote.value = False
    tx_filter.failed.value = False
    request.transactions["pump_fun"].CopyFrom(tx_filter)
    request.commitment = geyser_pb2.CommitmentLevel.CONFIRMED
    yield request

# Stream events
metadata = [("x-api-key", "YOUR_SUBGLOW_KEY")]
stream = stub.Subscribe(build_request(), metadata=metadata)

for update in stream:
    if update.HasField("transaction"):
        slot = update.transaction.slot
        sig = update.transaction.transaction.signature.hex()[:16]
        print(f"Pump.fun tx: {sig}... | Slot: {slot}")

Filter by Multiple Programs

Add Raydium alongside Pump.fun for comprehensive DEX coverage:

raydium_filter = geyser_pb2.SubscribeRequestFilterTransactions()
raydium_filter.account_include.append("675kPX9MHTjS2zt1qfr1NYHuzeLXfQM9H24wFSUt1Mp8")
raydium_filter.vote.value = False
raydium_filter.failed.value = False
request.transactions["raydium_amm"].CopyFrom(raydium_filter)

Handle Disconnection

import time

def stream_with_retry():
    delay = 1
    while True:
        try:
            channel = grpc.secure_channel("grpc.subglow.io:443", grpc.ssl_channel_credentials())
            stub = geyser_pb2_grpc.GeyserStub(channel)
            stream = stub.Subscribe(build_request(), metadata=metadata)
            for update in stream:
                delay = 1  # Reset on success
                process_update(update)
        except grpc.RpcError as e:
            print(f"Stream error: {e.code()}. Reconnecting in {delay}s...")
            time.sleep(delay)
            delay = min(delay * 2, 30)

Skip the Complexity with Subglow

Raw Yellowstone gRPC requires protobuf compilation, Borsh decoding, and instruction parsing. Subglow delivers the same data as pre-parsed JSON over gRPC. Your Python code gets clean dictionaries with human-readable field names — no protobuf stubs needed.

Solana gRPC Python Tutorial — Stream Blockchain Data with Yellowstone (2026)