From 2dd506e5830e685fec8dfe485c4019baf35ac94f Mon Sep 17 00:00:00 2001 From: bracesproul Date: Thu, 27 Jun 2024 16:54:59 -0700 Subject: [PATCH] cr --- docs/core_docs/docs/integrations/chat/bedrock.mdx | 2 +- docs/core_docs/docs/integrations/chat/bedrock_converse.mdx | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/docs/core_docs/docs/integrations/chat/bedrock.mdx b/docs/core_docs/docs/integrations/chat/bedrock.mdx index 7bd586142ea6..49c291da8fbc 100644 --- a/docs/core_docs/docs/integrations/chat/bedrock.mdx +++ b/docs/core_docs/docs/integrations/chat/bedrock.mdx @@ -10,7 +10,7 @@ sidebar_label: Bedrock ## Setup :::tip -The `ChatBedrockConverse` chat model is now available via `@langchain/aws`. Access tool calling with more models with this package. +The [`ChatBedrockConverse` chat model is now available via `@langchain/aws`](/docs/integrations/chat/bedrock_converse). Access tool calling with more models with this package. ::: You'll need to install the `@langchain/community` package: diff --git a/docs/core_docs/docs/integrations/chat/bedrock_converse.mdx b/docs/core_docs/docs/integrations/chat/bedrock_converse.mdx index 02781509912c..2b82a67898b6 100644 --- a/docs/core_docs/docs/integrations/chat/bedrock_converse.mdx +++ b/docs/core_docs/docs/integrations/chat/bedrock_converse.mdx @@ -50,7 +50,7 @@ import BedrockMultimodalExample from "@examples/models/chat/integration_bedrock_ See the LangSmith trace [here](https://smith.langchain.com/public/c40f8d09-123a-4b3b-934a-625d5ee0f57a/r). ::: -### Tool calling +## Tool calling The examples below demonstrate how to use tool calling, along with the `withStructuredOutput` method to easily compose structured output LLM calls. @@ -64,7 +64,7 @@ Check out the output of this tool call! We can see here it's using chain-of-thou See the LangSmith trace [here](https://smith.langchain.com/public/d34e378d-5044-4b5b-9ed7-3d2486fe5d47/r) ::: -#### `.withStructuredOutput({ ... })` +### `.withStructuredOutput({ ... })` Using the `.withStructuredOutput` method, you can easily make the LLM return structured output, given only a Zod or JSON schema: