This will help you get started with InMemoryStore. For detailed documentation of all InMemoryStore features and configurations head to the API reference. The InMemoryStore allows for a generic type to be assigned to the values in the store. We’ll assign type BaseMessage as the type of our values, keeping with the theme of a chat history store.

​
Overview

​
Integration details

ClassPackageLocalPY supportPackage downloadsPackage latest
InMemoryStore@langchain/core✅✅NPM - DownloadsNPM - Version

​
Setup

​
Installation

The LangChain InMemoryStore integration lives in the @langchain/core package: 

import IntegrationInstallTooltip from "@mdx_components/integration_install_tooltip.mdx";
<IntegrationInstallTooltip></IntegrationInstallTooltip>

<Npm2Yarn>
  @langchain/core
</Npm2Yarn>

​
Instantiation

Now we can instantiate our byte store: 
import { InMemoryStore } from "@langchain/core/stores"
import { BaseMessage } from "@langchain/core/messages";

const kvStore = new InMemoryStore<BaseMessage>();

​
Usage

You can set data under keys like this using the mset method: 
import { AIMessage, HumanMessage } from "@langchain/core/messages";

await kvStore.mset(
  [
    ["key1", new HumanMessage("value1")],
    ["key2", new AIMessage("value2")],
  ]
)

await kvStore.mget(
  [
    "key1",
    "key2",
  ]
)

[
  HumanMessage {
    "content": "value1",
    "additional_kwargs": {},
    "response_metadata": {}
  },
  AIMessage {
    "content": "value2",
    "additional_kwargs": {},
    "response_metadata": {},
    "tool_calls": [],
    "invalid_tool_calls": []
  }
]
And you can delete data using the mdelete method: 
await kvStore.mdelete(
  [
    "key1",
    "key2",
  ]
)

await kvStore.mget(
  [
    "key1",
    "key2",
  ]
)

[ undefined, undefined ]

​
Yielding values

If you want to get back all the keys you can call the yieldKeys method. Optionally, you can pass a key prefix to only get back keys which match that prefix. 
import { InMemoryStore } from "@langchain/core/stores"
import { AIMessage, BaseMessage, HumanMessage } from "@langchain/core/messages";

const kvStoreForYield = new InMemoryStore<BaseMessage>();

// Add some data to the store
await kvStoreForYield.mset(
  [
    ["message:id:key1", new HumanMessage("value1")],
    ["message:id:key2", new AIMessage("value2")],
  ]
)

const yieldedKeys = [];
for await (const key of kvStoreForYield.yieldKeys("message:id:")) {
  yieldedKeys.push(key);
}

console.log(yieldedKeys);

[ 'message:id:key1', 'message:id:key2' ]

​
API reference

For detailed documentation of all InMemoryStore features and configurations, head to the API reference