Building Subgraphs on Cosmos

This guide is an introduction on building subgraphs indexing Cosmos based blockchains.

What are Cosmos subgraphs?Link to this section

The Graph allows developers to process blockchain events and make the resulting data easily available via an open GraphQL API, known as a subgraph. Graph Node is now able to process Cosmos events, which means Cosmos developers can now build subgraphs to easily index on-chain events.

There are four types of handlers supported in Cosmos subgraphs:

  • Block handlers run whenever a new block is appended to the chain.
  • Event handlers run when a specific event is emitted.
  • Transaction handlers run when a transaction occurs.
  • Message handlers run when a specific message occurs.

Based on the official Cosmos documentation:

Events are objects that contain information about the execution of the application. They are mainly used by service providers like block explorers and wallets to track the execution of various messages and index transactions.

Transactions are objects created by end-users to trigger state changes in the application.

Messages are module-specific objects that trigger state transitions within the scope of the module they belong to.

Even though all data can be accessed with a block handler, other handlers enable subgraph developers to process data in a much more granular way.

Building a Cosmos subgraphLink to this section

Subgraph DependenciesLink to this section

graph-cli is a CLI tool to build and deploy subgraphs, version >=0.30.0 is required in order to work with Cosmos subgraphs.

graph-ts is a library of subgraph-specific types, version >=0.27.0 is required in order to work with Cosmos subgraphs.

Subgraph Main ComponentsLink to this section

There are three key parts when it comes to defining a subgraph:

subgraph.yaml: a YAML file containing the subgraph manifest, which identifies which events to track and how to process them.

schema.graphql: a GraphQL schema that defines what data is stored for your subgraph, and how to query it via GraphQL.

AssemblyScript Mappings: AssemblyScript code that translates from blockchain data to the entities defined in your schema.

Subgraph Manifest DefinitionLink to this section

The subgraph manifest (subgraph.yaml) identifies the data sources for the subgraph, the triggers of interest, and the functions (handlers) that should be run in response to those triggers. See below for an example subgraph manifest for a Cosmos subgraph:

specVersion: 0.0.5
description: Cosmos Subgraph Example
file: ./schema.graphql # link to the schema file
- kind: cosmos
name: CosmosHub
network: cosmoshub-4 # This will change for each cosmos-based blockchain. In this case, the example uses the Cosmos Hub mainnet.
startBlock: 0 # Required for Cosmos, set this to 0 to start indexing from chain genesis
apiVersion: 0.0.7
language: wasm/assemblyscript
- handler: handleNewBlock # the function name in the mapping file
- event: rewards # the type of the event that will be handled
handler: handleReward # the function name in the mapping file
- handler: handleTransaction # the function name in the mapping file
- message: /cosmos.staking.v1beta1.MsgDelegate # the type of a message
handler: handleMsgDelegate # the function name in the mapping file
file: ./src/mapping.ts # link to the file with the Assemblyscript mappings
  • Cosmos subgraphs introduce a new kind of data source (cosmos).
  • The network should correspond to a chain in the Cosmos ecosystem. In the example, the Cosmos Hub mainnet is used.

Schema DefinitionLink to this section

Schema definition describes the structure of the resulting subgraph database and the relationships between entities. This is agnostic of the original data source. There are more details on subgraph schema definition here.

AssemblyScript MappingsLink to this section

The handlers for processing events are written in AssemblyScript.

Cosmos indexing introduces Cosmos-specific data types to the AssemblyScript API.

class Block {
header: Header
evidence: EvidenceList
resultBeginBlock: ResponseBeginBlock
resultEndBlock: ResponseEndBlock
transactions: Array<TxResult>
validatorUpdates: Array<Validator>
class EventData {
event: Event
block: HeaderOnlyBlock
tx: TransactionContext
class TransactionData {
tx: TxResult
block: HeaderOnlyBlock
class MessageData {
message: Any
block: HeaderOnlyBlock
tx: TransactionContext
class TransactionContext {
hash: Bytes
index: u32
code: u32
gasWanted: i64
gasUsed: i64
class HeaderOnlyBlock {
header: Header
class Header {
version: Consensus
chainId: string
height: u64
time: Timestamp
lastBlockId: BlockID
lastCommitHash: Bytes
dataHash: Bytes
validatorsHash: Bytes
nextValidatorsHash: Bytes
consensusHash: Bytes
appHash: Bytes
lastResultsHash: Bytes
evidenceHash: Bytes
proposerAddress: Bytes
hash: Bytes
class TxResult {
height: u64
index: u32
tx: Tx
result: ResponseDeliverTx
hash: Bytes
class Event {
eventType: string
attributes: Array<EventAttribute>
class Any {
typeUrl: string
value: Bytes

Each handler type comes with its own data structure that is passed as an argument to a mapping function.

  • Block handlers receive the Block type.
  • Event handlers receive the EventData type.
  • Transaction handlers receive the TransactionData type.
  • Message handlers receive the MessageData type.

As a part of MessageData the message handler receives a transaction context, which contains the most important information about a transaction that encompasses a message. The transaction context is also available in the EventData type, but only when the corresponding event is associated with a transaction. Additionally, all handlers receive a reference to a block (HeaderOnlyBlock).

You can find the full list of types for the Cosmos integration here.

Message decodingLink to this section

It's important to note that Cosmos messages are chain-specific and they are passed to a subgraph in the form of a serialized Protocol Buffers payload. As a result, the message data needs to be decoded in a mapping function before it can be processed.

An example of how to decode message data in a subgraph can be found here.

Creating and building a Cosmos subgraphLink to this section

The first step before starting to write the subgraph mappings is to generate the type bindings based on the entities that have been defined in the subgraph schema file (schema.graphql). This will allow the mapping functions to create new objects of those types and save them to the store. This is done by using the codegen CLI command:

$ graph codegen

Once the mappings are ready, the subgraph needs to be built. This step will highlight any errors the manifest or the mappings might have. A subgraph needs to build successfully in order to be deployed to the Graph Node. It can be done using the build CLI command:

$ graph build

Deploying a Cosmos subgraphLink to this section

Once your subgraph has been created, you can deploy your subgraph by using the graph deploy CLI command after running the graph create CLI command:

Hosted Service

graph create account/subgraph-name --product hosted-service
graph deploy account/subgraph-name --product hosted-service

Local Graph Node (based on default configuration):

graph create subgraph-name --node http://localhost:8020
graph deploy subgraph-name --node http://localhost:8020/ --ipfs http://localhost:5001

Querying a Cosmos subgraphLink to this section

The GraphQL endpoint for Cosmos subgraphs is determined by the schema definition, with the existing API interface. Please visit the GraphQL API documentation for more information.

Supported Cosmos BlockchainsLink to this section

Cosmos HubLink to this section

What is Cosmos Hub?Link to this section

The Cosmos Hub blockchain is the first blockchain in the Cosmos ecosystem. You can visit the official documentation for more information.

NetworksLink to this section

Cosmos Hub mainnet is cosmoshub-4. Cosmos Hub current testnet is theta-testnet-001.
Other Cosmos Hub networks, i.e. cosmoshub-3, are halted, therefore no data is provided for them.

OsmosisLink to this section

Osmosis support in Graph Node and on the Hosted Service is in beta: please contact the graph team with any questions about building Osmosis subgraphs!

What is Osmosis?Link to this section

Osmosis is a decentralized, cross-chain automated market maker (AMM) protocol built on top of the Cosmos SDK. It allows users to create custom liquidity pools and trade IBC-enabled tokens. You can visit the official documentation for more information.

NetworksLink to this section

Osmosis mainnet is osmosis-1. Osmosis current testnet is osmo-test-4.

Example SubgraphsLink to this section

Here are some example subgraphs for reference:

Block Filtering Example

Validator Rewards Example

Validator Delegations Example

Osmosis Token Swaps Example

Edit page

Building Subgraphs on NEAR
Building Subgraphs on Arweave