```python
from langchain.chains import ConversationChain
from langchain.memory import ConversationBufferMemory

conversation = ConversationChain(
    llm=chat,
    memory=ConversationBufferMemory()
)

conversation.run("Answer briefly. What are the first 3 colors of a rainbow?")
# -> The first three colors of a rainbow are red, orange, and yellow.
conversation.run("And the next 4?")
# -> The next four colors of a rainbow are green, blue, indigo, and violet.
```

<CodeOutputBlock lang="python">

```
    'The next four colors of a rainbow are green, blue, indigo, and violet.'
```

</CodeOutputBlock>

Essentially, `BaseMemory` defines an interface of how `langchain` stores memory. It allows reading of stored data through `load_memory_variables` method and storing new data through `save_context` method. You can learn more about it in the [Memory](/docs/modules/memory/) section.