Skip to main content


AppData schema definitions

These schemas are used in the data encoded on appData field for CowProtocol orders.

For more details, check the docs.


yarn add @cowprotocol/app-data


import { MetadataApi } from '@cowprotocol/app-data'

export const metadataApi = new MetadataApi()

const appCode = 'YOUR_APP_CODE'
const environment = 'prod'
const referrer = { address: `REFERRER_ADDRESS` }

const quote = { slippageBips: '0.5' } // Slippage percent, it's 0 to 100
const orderClass = { orderClass: 'market' } // "market" | "limit" | "liquidity"

const appDataDoc = await metadataApi.generateAppDataDoc({
metadata: {

const { cid, appDataHex, appDataContent } = await metadataApi.appDataToCid(appDataDoc)

// 💡🐮 You should use appDataHex as the appData value in the CoW Order. "cid" Identifies the metadata associated to the CoW order in IPFS

// You can derive the CID from the appDataHex of any order
const actualCid = await metadataApi.appDataHexToCid(appDataHex)
console.log(cid === actualCid) // Should be true

// You can derive the appDataHex from the CID of any order
const actualAppDatahex = await metadataApi.appDataHexToCid(cid)
console.log(appDataHex === actualAppDatahex) // Should be true

// You can retrieve the JSON document from the CID
// 🔔 NOTE: for this to work, someone needs to upload the document to IPFS (the CoW API does it, but anyone could upload it too)
const actualAppDoc = await fetchDocFromCid(cid)

// You can also retrieve the JSON from the appDataHex
const actualAppDoc2 = await fetchDocFromAppDataHex(appDataHex)


Schemas are exposed as json files, where the version is the file name:

// Getting the version v0.4.0
const schema = require('@cowprotocol/app-data/schemas/v0.4.0.json')

// Now you can for example run validation against a schema

Type definitions

There are also type definitions

import { v0_4_0 } from '@cowprotocol/app-data'

// Note: this example is
function createAppDataV0_4_0(appCode: v0_4_0.AppCode, metadata: v0_4_0.Metadata): v0_4_0.AppDataRootSchema {
return {
version: '0.4.0',


The latest version names are exposed as constants

import {
} from '@cowprotocol/app-data'


Get appData schema

To get a schema definition by version

import { getAppDataSchema } from '@cowprotocol/app-data'

const schema = getAppDataSchema('0.1.0')

It'll throw if the version does not exist

Validate appDataDoc

To validate a document, pass it to validateAppDataDoc. It'll return an object with a boolean indicating success and errors, if any. The version to validate against will be taken from the doc itself.

import { validateAppDataDoc } from '@cowprotocol/app-data'

let doc = { version: '0.4.0', metadata: {} }

let result = await validateAppDataDoc(doc)
console.log(result) // { success: true }

doc = { version: '0.0.0', metadata: {} }

result = await validateAppDataDoc(doc)
// Contrary to `getAppDataSchema`, invalid or non-existing schemas won't throw
console.log(result) // { success: false, errors: 'AppData version 0.0.0 doesn\'t exist'}


Fork the repo so you can create a new PR. Then:

  1. Add a new version for the schema using the semver convention
  • Just duplicate the latest version i.e. src/schemas/<old-version>.json to src/schemas/<new-version>.json
  1. If you are adding a new meta-data
  1. If you are modifying an existing meta-data
  • Version it using the semver convention
  • You will need to create the new file for the meta-data schema: <meta-data-name>/<new-version>.json
  • Update it in the main schema you just created in step 1: Set it to "<meta-data-name>": { "$ref": "<meta-data-name>/<new-version>.json#" }
  1. Modify the compile.ts script
  1. Generate the typescript types
  • Run yarn build
  1. Make a test focusing on the new or modified meta-data:
  1. Create the PR and document it together with the motivation for the changes