This section enumerates what you need to do to write a client that communicates with lnd in Python.

Setup and Installation

Lnd uses the gRPC protocol for communication with clients like lncli. gRPC is based on protocol buffers and as such, you will need to compile the lnd proto file in Python before you can use it to communicate with lnd.

  1. Create a virtual environment for your project
     $ virtualenv lnd
  2. Activate the virtual environment
     $ source lnd/bin/activate
  3. Install dependencies (googleapis-common-protos is required due to the use of google/api/annotations.proto)
     (lnd)$ pip install grpcio grpcio-tools googleapis-common-protos
  4. Clone the google api’s repository (required due to the use of google/api/annotations.proto)
     (lnd)$ git clone
  5. Copy the lnd rpc.proto file (you’ll find this at lnrpc/rpc.proto) or just download it
     (lnd)$ curl -o rpc.proto -s
  6. Compile the proto file
     (lnd)$ python -m grpc_tools.protoc --proto_path=googleapis:. --python_out=. --grpc_python_out=. rpc.proto

After following these steps, two files and will be generated. These files will be imported in your project anytime you use Python gRPC.

Generating RPC modules for subservers

If you want to use any of the subservers’ functionality, you also need to generate the python modules for them.

For example, if you want to generate the RPC modules for the Router subserver (located/defined in routerrpc/router.proto), you need to run the following two extra steps (after completing all 6 step described above) to get the and

(lnd)$ curl -o router.proto -s
(lnd)$ python -m grpc_tools.protoc --proto_path=googleapis:. --python_out=. --grpc_python_out=. router.proto

Imports and Client

Every time you use Python gRPC, you will have to import the generated rpc modules and set up a channel and stub to your connect to your lnd node:

import rpc_pb2 as ln
import rpc_pb2_grpc as lnrpc
import grpc
import os

# Due to updated ECDSA generated tls.cert we need to let gprc know that
# we need to use that cipher suite otherwise there will be a handhsake
# error when we communicate with the lnd rpc server.

# Lnd cert is at ~/.lnd/tls.cert on Linux and
# ~/Library/Application Support/Lnd/tls.cert on Mac
cert = open(os.path.expanduser('~/.lnd/tls.cert'), 'rb').read()
creds = grpc.ssl_channel_credentials(cert)
channel = grpc.secure_channel('localhost:10009', creds)
stub = lnrpc.LightningStub(channel)


Let’s walk through some examples of Python gRPC clients. These examples assume that you have at least two lnd nodes running, the RPC location of one of which is at the default localhost:10009, with an open channel between the two nodes.

Simple RPC

# Retrieve and display the wallet balance
response = stub.WalletBalance(ln.WalletBalanceRequest())

Response-streaming RPC

request = ln.InvoiceSubscription()
for invoice in stub.SubscribeInvoices(request):

Now, create an invoice for your node at localhost:10009and send a payment to it from another node.

$ lncli addinvoice --amt=100
	"r_hash": <R_HASH>,
	"pay_req": <PAY_REQ>
$ lncli sendpayment --pay_req=<PAY_REQ>

Your Python console should now display the details of the recently satisfied invoice.

Bidirectional-streaming RPC

from time import sleep
import codecs

def request_generator(dest, amt):
      # Initialization code here
      counter = 0
      print("Starting up")
      while True:
          request = ln.SendRequest(
          yield request
          # Alter parameters here
          counter += 1

# Outputs from lncli are hex-encoded
dest_bytes = codecs.decode(dest_hex, 'hex')

request_iterable = request_generator(dest=dest_bytes, amt=100)

for payment in stub.SendPayment(request_iterable):

This example will send a payment of 100 satoshis every 2 seconds.

Using Macaroons

To authenticate using macaroons you need to include the macaroon in the metadata of the request.

import codecs

# Lnd admin macaroon is at ~/.lnd/data/chain/bitcoin/simnet/admin.macaroon on Linux and
# ~/Library/Application Support/Lnd/data/chain/bitcoin/simnet/admin.macaroon on Mac
with open(os.path.expanduser('~/.lnd/data/chain/bitcoin/simnet/admin.macaroon'), 'rb') as f:
    macaroon_bytes =
    macaroon = codecs.encode(macaroon_bytes, 'hex')

The simplest approach to use the macaroon is to include the metadata in each request as shown below.

stub.GetInfo(ln.GetInfoRequest(), metadata=[('macaroon', macaroon)])

However, this can get tiresome to do for each request, so to avoid explicitly including the macaroon we can update the credentials to include it automatically.

def metadata_callback(context, callback):
    # for more info see grpc docs
    callback([('macaroon', macaroon)], None)

# build ssl credentials using the cert the same as before
cert_creds = grpc.ssl_channel_credentials(cert)

# now build meta data credentials
auth_creds = grpc.metadata_call_credentials(metadata_callback)

# combine the cert credentials and the macaroon auth credentials
# such that every call is properly encrypted and authenticated
combined_creds = grpc.composite_channel_credentials(cert_creds, auth_creds)

# finally pass in the combined credentials when creating a channel
channel = grpc.secure_channel('localhost:10009', combined_creds)
stub = lnrpc.LightningStub(channel)

# now every call will be made with the macaroon already included


With the above, you should have all the lnd related gRPC dependencies installed locally into your virtual environment. In order to get up to speed with protofbuf usage from Python, see this official protobuf tutorial for Python. Additionally, this official gRPC resource provides more details around how to drive gRPC from Python.

API documentation

There is an online API documentation available that shows all currently existing RPC methods, including code snippets on how to use them.