# Interface
Remember the bdk-cli --help
command you ran before? Let's analyze its output here to figure out the interface:
# Flags
FLAGS:
-h, --help Prints help information
-V, --version Prints version information
These are the optional flags that can be set with every command. The -h
flag prints the help message, the -V
flag only prints the version.
# Verbosity
If you want to increase the verbosity of the output, you should use the RUST_LOG
environment variable. You can set it like so to see a lot more of what's going on behind the scenes, before running the bdk-cli
command. You only have to do this once when you open a new shell, after that you can run the bdk-cli
command multiple times.
export RUST_LOG="bdk=debug"
# Options
OPTIONS:
-n, --network <NETWORK> Sets the network [default: testnet]
These are the global options that can be set. They are pretty much like the flags, but they also take a value.
The --network
flag can be used to change the network. Right now only testnet
and regtest
are supported since the code is very much not production-ready yet.
# Subcommands
# key
Commands to generate, restore and derive HD keys.
To get more details about every single subcommand to the key
command, run bdk-cli key
. Under the key
command we have the following subcommands to quickly generate, restore and derive HD keys.
Command | Description |
---|---|
generate | Generates new random seed mnemonic phrase and corresponding master extended key |
restore | Restore a master extended key from seed backup mnemonic words |
derive | Derive a child key pair from a master extended key and a derivation path string (eg. "m/84'/1'/0'/0" or "m/84h/1h/0h/0") |
# generate
OPTIONS:
-p, --password <PASSWORD> Seed password
-e, --entropy <WORD_COUNT> Entropy level based on number of random seed mnemonic words [default: 24] [possible values: 12, 24]
To use this subcommand, run bdk-cli key generate
. This creates a BIP32 HD wallet master extended private key, a wallet fingerprint and BIP39 mnemonic words. The xprv can be used with key derive
to derive extended public keys.
# restore
OPTIONS:
-m, --mnemonic <MNEMONIC> Seed mnemonic words, must be quoted (eg. "word1 word2 ...")
-p, --password <PASSWORD> Seed password
To use this subcommand, run bdk-cli key restore -m <MNEMONIC>
. Format <MNEMONIC>
as shown in the options above. This recovers a BIP32 HD wallet seed from BIP39 mnemonic words. It outputs a wallet fingerprint and a master extended private key.
# derive
OPTIONS:
-p, --path <PATH> Path to use to derive extended public key from extended private key
-x, --xprv <XPRV> Extended private key to derive from
To derive child keys from an extended private key, run bdk-cli key derive -p <PATH> -x <XPRV>
.
# wallet
These are the main "functions" of the wallet. Most of them are pretty self explanatory, but we'll go over them quickly anyways. You can get more details about the subcommands by running bdk-cli wallet
. In addition, you can also get more details about every single command by running bdk-cli wallet <subcommand> --help
.
Command | Description |
---|---|
broadcast | Broadcasts a transaction to the network. Takes either a raw transaction or a PSBT to extract |
bump_fee | Bumps the fees of an RBF transaction |
combine_psbt | Combines multiple PSBTs into one |
create_tx | Creates a new unsigned tranasaction |
extract_psbt | Extracts a raw transaction from a PSBT |
finalize_psbt | Finalizes a psbt |
get_balance | Returns the current wallet balance |
get_new_address | Generates a new external address |
list_transactions | Lists all the incoming and outgoing transactions of the wallet |
list_unspent | Lists the available spendable UTXOs |
policies | Returns the available spending policies for the descriptor |
public_descriptor | Returns the public version of the wallet's descriptor(s) |
help | Prints this message or the help of the given subcommand(s) |
sign | Signs and tries to finalize a PSBT |
sync | Syncs with the chosen Electrum server |
# Options
OPTIONS:
-c, --change_descriptor <CHANGE_DESCRIPTOR> Sets the descriptor to use for internal addresses
--conc <CONCURRENCY>
Number of parallel requests sent to the esplora service (default: 4) [default: 4]
-d, --descriptor <DESCRIPTOR> Sets the descriptor to use for the external addresses
-s, --server <ESPLORA_URL>
Use the esplora server if given as parameter [default: https://blockstream.info/testnet/api/]
-p, --proxy <PROXY_ADDRS:PORT> Sets the SOCKS5 proxy for Blockchain backend
-r, --retries <PROXY_RETRIES> Sets the SOCKS5 proxy retries for the Electrum client [default: 5]
-a, --proxy_auth <PROXY_USER:PASSWD> Sets the SOCKS5 proxy credential
-g, --stop_gap <STOP_GAP>
Stop searching addresses for transactions after finding an unused gap of this length [default: 10]
--timeout <TIMEOUT> Socket timeout [default: 5]
-w, --wallet <WALLET_NAME> Selects the wallet to use
These are the global options that can be set. They are pretty much like the flags, but they also take a value. The only required one is the --descriptor
or -d
flag, since every wallet must have an
associated descriptor.
The --change-descriptor
flag can be used to set a different descriptor for the change addresses, sometimes called "internal" addresses in Bitcoin Core. Unfortunately there isn't
really consensus (opens new window) on a nice way to encode information about the change derivation inside the standard descriptor, so we are stuck with having two separate ones. Keep in mind though, that even if you don't specify a change descriptor, you'll still be able to create transactions - the change address will simply be generated from the standard descriptor.
The --server
flag can be used to select the Electrum server to use. By default it's connecting to Blockstream's electrum servers, which seems pretty stable.
If you are having connection issues, you can also try with one of the other servers listed here (opens new window) and see if you have more luck with those.
Right now both plaintext and ssl servers are supported (prefix tcp://
or no prefix at all for tcp, prefix ssl://
for ssl). This flag can also be used to connect to an Esplora server. It should be set to the API's "base url". For public instances of Esplora this is https://blockstream.info/api
for mainnet and https://blockstream.info/testnet/api
for testnet.
The --proxy
flag can be optionally used to specify a SOCKS5 proxy to use when connecting to an electrum
, esplora
or compact_filters
blockchain backend. Spawning a local Tor daemon and using it as a proxy will allow you to connect to the backend's .onion
URLs. Keep in mind that only plaintext server are supported over a proxy.
The --retries
flag can be optionally used to specify the number of retries when connecting to
an electrum
, esplora
or compact_filters
blockchain backend via the SOCKS5 proxy.
The --proxy_auth
flag can be optionally used to set the SOCKS5 proxy authentication credentials.
The --stop_gap
flag can be optionally used to stop searching addresses for transactions after finding an unused gap of this length.
The --wallet
flag can be used to select which wallet to use, if you have more than one of them.
If no --wallet
option is passed, a random wallet name is generated based on the hash of the descriptor string. If you get a ChecksumMismatch
error when you make some changes to your descriptor, it's because it does not match the one you've used to initialize the cache. One solution could be to switch to a new wallet name, or delete the cache directory at ~/.bdk-bitcoin
and start from scratch.
# Subcommands
# broadcast
OPTIONS:
--psbt <BASE64_PSBT> Sets the PSBT to extract and broadcast
--tx <RAWTX> Sets the raw transaction to broadcast
Broadcasts a transaction. The transaction can be a raw hex transaction or a PSBT, in which case it also has to be "finalizable" (i.e. it should contain enough partial signatures to construct a finalized valid scriptsig/witness).
# bump_fee
FLAGS:
--offline_signer
Make a PSBT that can be signed by offline signers and hardware wallets. Forces the addition of `non_witness_utxo` and more details to let the signer identify the change output
OPTIONS:
--unspendable <CANT_SPEND_TXID:VOUT>...
Marks an utxo as unspendable, in case more inputs are needed to cover the extra fees
--utxos <MUST_SPEND_TXID:VOUT>...
Selects which utxos *must* be added to the tx. Unconfirmed utxos cannot be used
-f, --fee_rate <SATS_VBYTE> The new targeted fee rate in sat/vbyte
-s, --shrink <SHRINK_ADDRESS>
Allows the wallet to reduce the amount to the specified address in order to increase fees
-t, --txid <TXID> TXID of the transaction to update
Bumps the fee of a transaction made with RBF. The transaction to bump is specified using the --txid
flag and the new fee rate with --fee_rate
.
# combine_psbt
OPTIONS:
--psbt <BASE64_PSBT>... Add one PSBT to comine. This option can be repeated multiple times, one for each PSBT
Combines multiple PSBTs by merging metadata and partial signatures. It can be used to merge multiple signed PSBTs into a single PSBT that contains every signature and is ready to be finalized.
# create_tx
FLAGS:
-r, --enable_rbf Enables Replace-By-Fee (BIP125)
--offline_signer Make a PSBT that can be signed by offline signers and hardware wallets. Forces the addition
of `non_witness_utxo` and more details to let the signer identify the change output
-a, --send_all Sends all the funds (or all the selected utxos). Requires only one recipients of value 0
OPTIONS:
--to <ADDRESS:SAT>... Adds a recipient to the transaction
--unspendable <CANT_SPEND_TXID:VOUT>... Marks a utxo as unspendable
--external_policy <EXT_POLICY>
Selects which policy should be used to satisfy the external descriptor
--internal_policy <INT_POLICY>
Selects which policy should be used to satisfy the internal descriptor
--utxos <MUST_SPEND_TXID:VOUT>... Selects which utxos *must* be spent
-f, --fee_rate <SATS_VBYTE> Fee rate to use in sat/vbyte
Creates a new unsigned PSBT. The flags allow to set a custom fee rate (the default is 1.0 sat/vbyte) with --fee_rate
or -f
, the list of UTXOs that should be considered unspendable with --unspendable
(this option can be specified multiple times) and a list of UTXOs that must be spent with --utxos
(again, this option can also be specified multiple times).
The --to
option sets the receiver address of the transaction, and should contain the address and amount in Satoshi separated by a colon, like: --to 2NErbQPsooXRatRJdrXDm9wKR2fRiZDT9wL:50000
. This option can also be specified multiple times to send to multiple addresses at once.
The --send_all
flag can be used to send the value of all the spendable UTXOs to a single address, without creating a change output. If this flag is set, there must be only one --to
address, and its value will be ignored (it can be set to 0).
The --external_policy
and --internal_policy
options are two advanced flags that can be used to select the spending policy that the sender intends to satisfy in this transaction. They are normally not required if there's no ambiguity, but sometimes with a complex descriptor one or both of them have to be specified, or you'll get a SpendingPolicyRequired
error. Those flags should be set to a JSON object that maps a policy node id to the list of child indexes that
the user intends to satisfy for that node. This is probably better explained with an example:
Let's assume our descriptor is: sh(thresh(2,pk(A),sj:and_v(v:pk(B),n:older(6)),snj:and_v(v:pk(C),after(630000))))
. There are three conditions and we need to satisfy two of them to be able to spend. The conditions are:
- Sign with the key corresponding to
pk(A)
- Sign with the key corresponding to
pk(B)
AND wait 6 blocks - Sign with the key corresponding to
pk(C)
AND wait that block 630,000 is reached
So if we write down all the possible outcomes when we combine them, we get:
- Sign with
pk(A)
+pk(B)
+ wait 6 blocks - Sign with
pk(A)
+pk(C)
+ wait block 630,000 - Sign with
pk(B)
+pk(C)
+ wait 6 blocks + wait block 630,000
In other words:
- If we choose option #1, the final transaction will need to have the
nSequence
of its inputs set to a value greather than or equal to 6, but thenLockTime
can stay at 0. - If we choose option #2, the final transaction will need to have its
nLockTime
set to a value greater than or equal to 630,000, but thenSequence
can be set to a final value. - If we choose option #3, both the
nSequence
andnLockTime
must be set.
The wallet can't choose by itself which one of these combination to use, so the user has to provide this information with the --external_policy
flag.
Now, let's draw the condition tree to understand better how the chosen policy is represented: every node has its id shown right next to its name, like qd3um656
for the root node. These ids can be seen by running the policies command.
Some ids have been omitted since they are not particularly relevant, in this example we will actually only use the root id.
Let's imagine that we are walking down from the root, and we want to use option #1. So we will have to select pk(A)
+ the whole B
node. Since these nodes have an id, we can use it to refer to them and say which children
we want to use. In this case we want to use children #0 and #1 of the root, so our final policy will be: --external_policy {"qd3um656":[0,1]}
.
# extract_psbt
OPTIONS:
--psbt <BASE64_PSBT> Sets the PSBT to extract
Extracts the global transaction from a PSBT. Note that partial signatures are ignored in this step. If you want to merge the partial signatures back into the global transaction first, please use finalize_psbt first
# finalize_psbt
OPTIONS:
--psbt <BASE64_PSBT> Sets the PSBT to finalize
--assume_height <HEIGHT> Assume the blockchain has reached a specific height
--trust_witness_utxo <WITNESS> Whether the signer should trust the witness_utxo, if the non_witness_utxo hasn’t been provided
Tries to finalize a PSBT by merging all the partial signatures and other elements back into the global transaction. This command fails if there are timelocks that have not yet expired, but the check can be overridden by specifying --assume_height
to make the wallet assume that a future height has already been reached.
# get_balance
This subcommand has no extra flags, and simply returns the available balance in Satoshis. This command should normally be called after sync
, since it only looks into the local cache to determine the list of UTXOs.
# get_new_address
This subcommand has no extra flags and returns a new address. It internally increments the derivation index and saves it in the database.
# list_transactions
This subcommand has no extra flags and returns the history of transactions made or received by the wallet, with their txid, confirmation height and the amounts (in Satoshi) "sent" (meaning, the sum of the wallet's inputs spent in the transaction) and
"received" (meaning, the sum of the outputs received by the wallet). Just like get_balance
it should normally be called after sync
, since it only operates
on the internal cache.
# list_unspent
This subcommand has no extra flags and returns the list of available UTXOs and their value in Satoshi. Just like get_balance
it should normally be called after sync
, since it only operates
on the internal cache.
# policies
This subcommand has no extra flags and returns the spending policies encoded by the descriptor in a more human-readable format. As an example, running the policies
command on the descriptor shown earlier for the explanation of the create_tx command, it will return this:
{
"id":"qd3um656",
"type":"THRESH",
"items":[
{
"id":"ykfuwzkl",
"type":"SIGNATURE",
"pubkey":"...",
"satisfaction":{
"type":"NONE"
},
"contribution":{
"type":"COMPLETE",
"condition":{
}
}
},
{
"id":"ms3xjley",
"type":"THRESH",
"items":[
{
"id":"xgfnp9rt",
"type":"SIGNATURE",
"pubkey":"...",
"satisfaction":{
"type":"NONE"
},
"contribution":{
"type":"COMPLETE",
"condition":{
}
}
},
{
"id":"5j96ludf",
"type":"RELATIVETIMELOCK",
"value":6,
"satisfaction":{
"type":"NONE"
},
"contribution":{
"type":"COMPLETE",
"condition":{
"csv":6
}
}
}
],
"threshold":2,
"satisfaction":{
"type":"NONE"
},
"contribution":{
"type":"PARTIALCOMPLETE",
"n":2,
"m":2,
"items":[
0,
1
],
"conditions":{
"[0, 1]":[
{
"csv":6
}
]
}
}
},
{
"id":"d8jph6ax",
"type":"THRESH",
"items":[
{
"id":"gdl039m6",
"type":"SIGNATURE",
"pubkey":"...",
"satisfaction":{
"type":"NONE"
},
"contribution":{
"type":"COMPLETE",
"condition":{
}
}
},
{
"id":"xpf2twg8",
"type":"ABSOLUTETIMELOCK",
"value":630000,
"satisfaction":{
"type":"NONE"
},
"contribution":{
"type":"COMPLETE",
"condition":{
"timelock":630000
}
}
}
],
"threshold":2,
"satisfaction":{
"type":"NONE"
},
"contribution":{
"type":"PARTIALCOMPLETE",
"n":2,
"m":2,
"items":[
0,
1
],
"conditions":{
"[0, 1]":[
{
"timelock":630000
}
]
}
}
}
],
"threshold":2,
"satisfaction":{
"type":"NONE"
},
"contribution":{
"type":"PARTIALCOMPLETE",
"n":3,
"m":2,
"items":[
0,
1,
2
],
"conditions":{
"[0, 1]":[
{
"csv":6
}
],
"[0, 2]":[
{
"timelock":630000
}
],
"[1, 2]":[
{
"csv":6,
"timelock":630000
}
]
}
}
}
This is a tree-like recursive structure, so it tends to get huge as more and more pieces are added, but it's in fact fairly simple. Let's analyze a simple node of the tree:
{
"id":"qd3um656",
"type":"SIGNATURE",
"pubkey":"...",
"satisfaction":{
"type":"NONE"
},
"contribution":{
"type":"COMPLETE",
"condition":{}
}
}
id
is a unique identifier to this specific node in the tree.type
, as the name implies, represents the type of node. It defines what should be provided to satisfy that particular node. Generally some other data are provided to give meaning to the type itself (like thepubkey
field here in the example). There are basically two families of types: some of them can only be used as leaves, while some other can only be used as intermediate nodes.Possible leaf nodes are:
SIGNATURE
, requires a signature made with the specified key. Has apubkey
if it's a single key, afingerprint
if the key is an xpub, or apubkey_hash
if the full public key is not present in the descriptor.SIGNATUREKEY
, requires a signature plus the raw public key. Again, it can have apubkey
,fingerprint
orpubkey_hash
.SHA256PREIMAGE
, requires the preimage of a givenhash
.HASH256PREIMAGE
, requires the preimage of a givenhash
.RIPEMD160PREIMAGE
, requires the preimage of a givenhash
.HASH160PREIMAGE
, requires the preimage of a givenhash
.ABSOLUTETIMELOCK
, doesn't technically require anything to be satisfied, just waiting for the timelock to expire. Has avalue
field with the raw value of the timelock (can be both in blocks or time-based).RELATIVETIMELOCK
, again only requires waiting for the timelock to expire. Has avalue
likeABSOLUTETIMELOCK
.
Possible non-leaf nodes are:
THRESH
, defines a threshold of policies that has to be met to satisfy the node. Has anitems
field, which is a list of policies to satisfy and athreshold
field that defines the threshold.MULTISIG
, Similar toTHRESH
, has akeys
field, which is a list of keys represented again as eitherpubkey
,fingerprint
orpubkey_hash
and athreshold
field.
satisfaction
is currently not implemented and will be used to provide PSBT introspection, like understanding whether or not a node is already satisfied and to which extent in a PSBT.contribution
represents if so and how much, the provided descriptor can contribute to the node.The possible types are:
NONE
, which means that the descriptor cannot contribute.COMPLETE
, which means that the descriptor by itself is enough to completely satisfy the node. It also adds acondition
field which represent any potential extra condition that has to be met to consider the node complete. An example are the timelock nodes, that are always complete but they have an extracsv
ortimelock
condition.PARTIAL
, which means that the descriptor can partially satisfy the policy. This adds am
,n
,items
that respectively represent the threshold, the number of available items to satisfy and the items that the provided descriptor can satisfy. Also adds aconditions
field which is an integer to list of conditions map. The key is the child index and the map are all the possibile extra conditions that have to be satisfied if that node is used in the threshold. For instance, if you have a threshold of a SIGNATURE and a RELATIVETIMELOCK, in this order, theconditions
field will be1 ⇒ csv(x)
, because the item at index 1 needs the extra csv condition.PARTIALCOMPLETE
, which is basically aPARTIAL
with the size ofitems
>=m
. It's treated as a separate entity to make the code a bit more clean and easier to implement. LikePARTIAL
, it also has am
,n
,items
fields but theconditions field
is a bit different: it's a list of integers to list of conditions map. The key represents the combination that can be used to satisfy the threshold, and the value contains all the possible conditions that also have to be satisfied. For instance, if you have a 2-of-2 threshold of a TIMELOCK and a RELATIVETIMELOCK, theconditions
field will be[0, 1] ⇒ csv(x) + timelock(y)
, because if the combination of items 0 and 1 is picked, both of their conditions will have to be meet too.
While the structure contains all of the intermediate nodes too, the root node is the most important one because it defines how the descriptor can contribute to spend outputs sent to its addresses.
For instance, looking at the root node of the previous example (with the internal items
omitted) from a descriptor that has all the three private keys for keys A, B and C, we can clearly see that it can satisfy
the descriptor (type = PARTIALCOMPLETE
) and the three options are [0, 1] ⇒ csv(6)
(Option #1), [0, 2] ⇒ timelock(630,000)
(Option #2) or [1, 2] ⇒ csv(6) + timelock(630,000)
(Option #3).
{
"type":"THRESH",
"items":[],
"threshold":2,
"satisfaction":{
"type":"NONE"
},
"contribution":{
"type":"PARTIALCOMPLETE",
"n":3,
"m":2,
"items":[
0,
1,
2
],
"conditions":{
"[0, 1]":[
{
"csv":6
}
],
"[0, 2]":[
{
"timelock":630000
}
],
"[1, 2]":[
{
"csv":6,
"timelock":630000
}
]
}
}
}
# public_descriptor
This subcommand has no extra flags and returns the "public" version of the wallet's descriptor(s). It can be used to bootstrap a watch-only instance for the wallet.
# help
This subcommand has no extra flags. When used as bdk-cli wallet help
it will displays all the
options and subcommands of the wallet
command. When used as bdk-cli wallet help <subcommand>
it will display all the details of the subcommand.
# sign
OPTIONS:
--psbt <BASE64_PSBT> Sets the PSBT to sign
--assume_height <HEIGHT> Assume the blockchain has reached a specific height. This affects the transaction finalization, if there are timelocks in the descriptor
--trust_witness_utxo <WITNESS> Whether the signer should trust the witness_utxo, if the non_witness_utxo hasn’t been provided
Adds to the PSBT all the signatures it can produce with the secrets embedded in the descriptor (xprv or WIF keys). Returns the signed PSBT and, if there are enough item to satisfy the script, also the extracted raw Bitcoin transaction.
Optionally, the --assume_height
option can be specified to let the wallet assume the blockchain has reached a specific height. This affects the finalization of the PSBT which is done right at the end of the signing process: the wallet tries to satisfy the spending condition of each input using the partial signatures collected. In case timelocks are present the wallet needs to know whether or not they have expired. This flag is particularly useful for offline wallets.
# sync
This subcommand has no extra flags. It connects to the chosen Electrum server and synchronizes the list of transactions received and available UTXOs.