This repository contains all specification files used to generate code for the Rosetta API.
The Rosetta API is specified in the OpenAPI 3.0 format (the successor to Swagger/OpenAPI 2.0). Requests and responses can be crafted with auto-generated code using Swagger Codegen or OpenAPI Generator, are human-readable (easy to debug and understand), and can be used in servers and browsers.
Before diving into the specification and documentation, we recommend reviewing the high-level diagrams below to gain some familiarity with the functionality of the Rosetta APIs.
The Rosetta APIs are organized into two distinct categories, the Data API and the Construction API. Simply put, the Data API is for retrieving data from a blockchain network and the Construction API is for constructing and submitting transactions to a blockchain network.
The Data API was previously known as the Node API and the Construction API was previously known as the Wallet API. Their names were changed to better reflect their functionality.
Caller (i.e. Coinbase) + Data API Implementation
+------------------------------------------------------------------------------------+
X |
X Get Supported Networks +---------------------------------> /network/list
X | +
X +-----------+--------------------------------------------------------+
Get supported networks, X | v |
their supported options, X | Get Network Options +------------------------------------> /network/options
and their status X | |
X +-----------+ |
X v |
X Get Network Status +-------------------------------------> /network/status
X |
|
X |
X Get Block +----------------------------------------------> /block
X | +
+---------+X +-----------------------------------------------+
| X v |
Ensure balance computed | X [Optional] Get Additional Block Transactions +-----------> /block/transaction
from block operations | X |
equals balance on node | |
| |
+-----------> Get account balance for each +---------------------------> /account/balance
account seen in a block |
|
|
X |
X Get Mempool Transactions +-------------------------------> /mempool
Monitor the mempool for X | +
broadcast transaction status X +-----------------------------------------------+
and incoming deposits X v |
X Get a Specific Mempool Transaction +---------------------> /mempool/transaction
X |
|
|
X |
Make arbitrary call to to access X Make Arbitrary Procedure Call +--------------------------> /call
network-specific data X |
X |
+
If you have seen illustrations of the flow of transaction construction in other blockchain SDKs, the following diagrams may seem peculiar to you. Unlike traditional SDKs, Construction API implementations are fully stateless, can perform construction offline (when metadata like an account nonce is pre-fetched), and never have access to private key material.
Caller (i.e. Coinbase) + Construction API Implementation
+-------------------------------------------------------------------------------------------+
|
Derive Address +----------------------------> /construction/derive
from Public Key |
|
X |
X Create Metadata Request +---------------------> /construction/preprocess
X (array of operations) | +
Get metadata needed X | |
to construct transaction X +-----------------------------------------------+
X v |
X Fetch Online Metadata +-----------------------> /construction/metadata (online)
X |
|
X |
X Construct Payloads to Sign +------------------> /construction/payloads
X (array of operations) | +
X | |
Create unsigned transaction X +------------------------------------------------+
X v |
X Parse Unsigned Transaction +------------------> /construction/parse
X to Confirm Correctness |
X |
|
X |
X Sign Payload(s) +-----------------------------> /construction/combine
X (using caller's own detached signer) | +
X | |
Create signed transaction X +-----------------------------------------------+
X v |
X Parse Signed Transaction +--------------------> /construction/parse
X to Confirm Correctness |
X |
|
X |
X Get hash of signed transaction +--------------> /construction/hash
Broadcast Signed Transaction X to monitor status |
X |
X Submit Transaction +--------------------------> /construction/submit (online)
X |
+
Many developers may not have security constraints that dictate construction must occur offline or that they use their own detached signer (making the previously described flow much more cumbersome than other transaction construction and signing SDKs).
Fortunately, it is possible (and encouraged) to build higher-level interfaces on top of these low-level endpoints to simplify development for integrators. For example, an interface developer may wish to automatically fetch metdata during their call to construct a transaction so users would not even know there are multiple interactions occurring. One could also provide a signing library with their higher-level interface so users do not need to use a detached signer. Here is an example of a simplified flow:
Caller (i.e. Mobile Wallet) + Construction API Implementation
+-------------------------------------------------------------------------------------------+
|
Derive Address +----------------------------> /construction/derive
from Public Key |
| X
| X Fetch metadata needed for
| X construction automatically
X | X
Create unsigned transaction X Construct Payloads to Sign +------------------> /construction/payloads X /construction/preprocess
X (array of operations) | X +
X | X v
| X /construction/metadata
X | X
X Sign Payload(s) +-----------------------------> /construction/combine
X (using rosetta-sdk-go keys package) | +
X | |
Create signed transaction X +-----------------------------------------------+
X v |
X Parse Signed Transaction +--------------------> /construction/parse
X to Confirm Correctness |
X |
|
X |
X Get hash of signed transaction +--------------> /construction/hash
Broadcast Signed Transaction X to monitor status |
X |
X Submit Transaction +--------------------------> /construction/submit
X |
+
Why go through all this trouble to build a higher-level interface on this Construction API instead of just using existing SDKs, you may be wondering? Any interface built on top of the Construction API could support the construction of transactions on any blockchain that supports Rosetta with no modification. You could, for example, build a WalletLink service that worked with any blockchain.
For blockchains with smart contracts, it is usually not possible to add high fidelity support for each deployed contract to the "core" Rosetta implementation. Accessing any contract often requires generating some SDK and making very contract-specific access decisions that can introduce significant complexity into the "core" implementation.
So, no smart contract interaction in Rosetta? Think again!
To empower developers to build Rosetta API implementations for a given smart contract
on top of your "core" implementation, we recommend implementing the /call
endpoint
and providing support for common network-specific endpoints that would be used to
read or modify contract state. In theory, you could stack any number of Rosetta API
implementations on top of each other where each child implementation calls some
set of primitives exposed in this /call
endpoint.
In the case of Ethereum, a group of developers could implement a core
rosetta-ethereum
implementation that provides the ability to track
and create ETH transfers and another group of developers could write a
rosetta-erc20
that allows for tracking and creating ERC-20 token transfers
using the eth_call
method exposed by rosetta-ethereum
. Another group of
developers could write an implementation that tracks DEX trades, lending activity,
or even validator performance. You can see an example of how this would work in
this illustration:
For some developers, the existing Data API and Construction API endpoints are not sufficient to fully support an asset integration. It is not possible, for example, to search for a transaction by hash or access all transactions that affected a particular account. Traditionally database-intensive functionality was purposely excluded from the collection of endpoints any Rosetta implementation must complete as to avoid imposing cumbersome requirements on Rosetta implementers (that could require maintaining architecture alongside their core node).
Because of the standardization introduced by Rosetta, it is possible to write a generic indexer for any Rosetta implementation. To avoid a proliferation of interfaces that service Rosetta in this layer, we've defined a set of standard "indexer" endpoints that enable developers to automatically integrate (with the SDK they already use to access the Rosetta API). Rosetta implementations are not required to implement "indexer" endpoints but are welcome to do so!
Indexer implementations must proxy non-indexer Data API and
Construction API calls to the implementation of interest (potentially
caching some data) so that developers do not need to connect to multiple
endpoints to access Rosetta. All calls contain a NetworkIdentifier
so it
should be possible to route requests without too much difficulty.
- Data API (proxied)
- Construction API (proxied)
/events/*
/search/*
If you think an endpoint is missing from this list, please reach out on our community.
Now that you have some familiarity with the flow of operations, we recommend taking a look at the Rosetta API Docs:
If you have any questions, don't hesitate to reach out in our community forum.
If you've made it this far, you are interested in developing a Rosetta API implementation for a blockchain project you are working on. As mentioned in the Rosetta doumentation, there is no restriction on what language you choose to use for your implementation and no repository where you must open a PR to share your work.
When you've finished an implementation for a blockchain, share your work in the ecosystem category of the community site. Platforms looking for implementations for certain blockchains will be monitoring this section of the website for high-quality implementations they can use for integration (make sure your implementation meets the "expectations" of any implementation).
For many developers, the best way to get started is to look at an example. For this reason, we developed a complete Rosetta API reference implementation for Bitcoin and Ethereum.
Developers of Bitcoin-like or Ethereum-like blockchains may find it easier to fork these reference implementations than to write an implementation from scratch.
If you are comfortable with Golang, the easiest way to write a Rosetta Data API implementation is to use rosetta-sdk-go. This Golang project provides a server package that empowers a developer to write a full Rosetta Data API server by only implementing an interface. This package automatically validates client requests and calls the functions you implement with pre-parsed requests (instead of in raw JSON).
There is a simple example of how to write an implementation using this package in rosetta-sdk-go.
If you plan to use a language other than Golang, you will need to either codegen a server (the overview mentions some tools that can help with this) or write one from scratch. If you do choose to write an implementation in another language, we ask that you create a separate repository in an SDK-like format for all the code you generate so that other developers can use it (see the note on SDKs in more languages). Check out rosetta-sdk-go for an example of how to generate code from this specification.
The Construction API only supports detached, curve-based key generation and signing. In short, this means that Construction API implementations will never have access to private keys and will only ever interact with public keys and signatures in standardized formats. These formats are listed below:
CurveType and SignatureType are purposely decoupled as a curve could be used
with multiple signature types (i.e. secp256k1:ECDSA
and secp256k1:Schnorr
).
- secp256k1: SEC compressed -
33 bytes
(https://secg.org/sec1-v2.pdf#subsubsection.2.3.3) - secp256r1: SEC compressed -
33 bytes
(https://secg.org/sec1-v2.pdf#subsubsection.2.3.3) - edwards25519:
y (255-bits) || x-sign-bit (1-bit)
-32 bytes
(https://ed25519.cr.yp.to/ed25519-20110926.pdf) - tweedle: 1st pk : Fq.t (32 bytes) || 2nd pk : Fq.t (32 bytes) (https://github.com/CodaProtocol/coda/blob/develop/rfcs/0038-rosetta-construction-api.md#marshal-keys)
- ecdsa:
r (32-bytes) || s (32-bytes)
-64 bytes
- ecdsa_recovery:
r (32-bytes) || s (32-bytes) || v (1-byte)
-65 bytes
- ed25519:
R (32-byte) || s (32-bytes)
-64 bytes
- schnorr_1:
r (32-bytes) || s (32-bytes)
-64 bytes
- schnorr_poseidon:
r (32-bytes) || s (32-bytes) where s = Hash(1st pk || 2nd pk || r)
-64 bytes
schnorr_1
is a EC-Schnorr signature implemented by Zilliqa where both r
and s
are scalars encoded as 32-bytes
values, most significant byte first. Refer to Zilliqa's Schnorr Library and Zilliqa Technical Whitepaper - Appendix A: Schnorr Digital Signature for details.)
schnorr_poseidon
is an EC-schnorr signature with Poseidon hash function implemented by O(1) Labs where both r
and s
are scalars encoded as 32-bytes
little-endian values. Refer to Coda's signer reference implementation
Unlike the Data API where there are no global type constraints (ex: you can specify any operation type), the Construction API has a clearly enumerated list of supported curves and signatures. Each one of these items has a clearly specified format that all implementations should expect to receive.
If a curve or signature you are employing is not enumerated in the specification, you will need to open a PR against the specification to add it along with the standard format it will be represented in.
It is up to the caller of a Construction API implementation to implement key generation
and signing for a particular CurveType:SignatureType. There is a keys
package
in rosetta-sdk-go that is commonly used by callers and anyone
in the community can implement additional schemes there.
Although the Construction API is defined in the same interface as endpoints that
are "online" (i.e. fetching a block with /block
), it must be possible to deploy your Data API
and Construction API separately. This does not mean implementations need to live
in separate repositories or even be defined in separate Dockerfiles. However, it must be possible
to start an "offline-only" version of your implementation that supports all Construction API
endpoints (other than /construction/metadata
and /construction/submit
).
/network/*
/block/*
/account/*
/mempool/*
/call
/construction/metadata
/construction/submit
/network/list
/network/options
/construction/derive
/construction/preprocess
/construction/payloads
/construction/combine
/construction/parse
/construction/hash
To validate your implementation, you'll need to run the rosetta-cli.
The rosetta-cli
has a collection of "check" commands that are used to ensure your implementation adheres to the
Rosetta API. check:data
ensures your implementation meets Data API requirements
and check:construction
ensures your implementation meets Construction API requirements.
You can view an extensive list of checks this tool performs here. If you'd like to add more checks for correctness, feel free to create an issue listing in detail what you think should be checked in any implementation.
Before starting your own integration, we recommend reading this blog post that walks through how to fetch blocks from a Ethereum Rosetta API implementation.
The most developed tools for working with Rosetta API implementations can
be found in rosetta-sdk-go. Below,
we highlight 2 packages in this repository that are very useful for parsing
data from a blockchain. You can find code examples of these packages throughout
the rosetta-cli, which uses rosetta-sdk-go
for all block processing.
The core of any integration is syncing blocks reliably. The syncer serially processes blocks from a Data API implementation (automatically handling re-orgs) with user-defined handling logic and pluggable storage. After a block is processed, store it to a DB or send a push notification...it's up to you!
When reading the operations in a block, it can be helpful to apply higher-level groupings to related operations or match operations in a transaction to some set of generic descriptions (i.e. ensure there are 2 operations of equal but opposite amounts). The parser empowers any integrator to build abstractions on top of the low-level building blocks that the Rosetta API exposes.
At Coinbase, we write a lot of our code in Golang. We knew we could build a great SDK in Golang so we put all of our effort into developing the rosetta-sdk-go.
We look forward to working with the community to develop just as powerful of an SDK in both Rust and JavaScript/TypeScript. If you'd like to work on SDKs in either of these languages, let us know on the community forum. We may feature a few of completed SDKs on the website or on this README.
make deps
to install dependenciesmake gen
to generate the specification filesmake check-valid
to ensure specification is validmake release
to check if code passes all tests run by CircleCI
This project is available open source under the terms of the Apache 2.0 License.
© 2020 Coinbase