# Add Support for Anthropic's Built-in Tools

[bleafman]

Add Support for Anthropic's Built-in Tools

Adds support for Anthropic's built-in tools (web search, text editor, bash) to bring feature parity with the Python implementation.

• Added support for server_tool_use, web_search_tool_result, and web_search_result content blocks
• Extended message formatting to handle server tool content blocks
• Added test coverage

Fixes #8215

Twitter Handle: @BrandonLeafman

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
langchainjs-docs	✅ Ready (Inspect)	Visit Preview		May 28, 2025 7:28pm

1 Skipped Deployment

Name	Status	Preview	Comments	Updated (UTC)
langchainjs-api-refs	⬜️ Ignored (Inspect)			May 28, 2025 7:28pm

[bleafman]

Here's some more background / rational for the changes:

Technical Details

Anthropic's built-in tools work differently from regular tools - they execute server-side and return specialized content blocks in API responses:

• server_tool_use: When Claude uses a built-in tool (like web search)
• web_search_tool_result: Container for web search results
• web_search_result: Individual search result within the container

These content blocks need to be parsed otherwise you get a "Unsupported message content format" errors on subsequent turns after the built-in tool usage.

The @anthropic-ai/sdk package needed to be bumped to access the WebSearchTool20250305 type.

Changes made:

1. types.ts: Added imports for the new content block types from @anthropic-ai/sdk
2. message_inputs.ts:
   • Added the new types to toolTypes array in _formatContent() (line ~294)
   • Updated normalizeContent() union type to include server tool types (line ~578)
3. message_outputs.ts: Enhanced streaming to handle server tool content blocks

I attempted to follows existing patterns - server tool content blocks get the same processing as regular tool blocks (index removal, input validation, etc.).

Python parity: The Python implementation already supported these via the Anthropic SDK's native types. This brings the JS version to the same level.

Testing: Added both unit tests (content block parsing) and integration tests (real API calls) to ensure reliability.

[dikaioai]

Please merge this!

[benjamincburns]

Thanks for submitting this, @bleafman! I had some questions, and a few requested changes. Hopefully it's nothing too troublesome, but let me know if you need me to clarify anything.

[benjamincburns]

Is there a union for built-in tools? If not, it would be good to create one in types.ts.

[bleafman]

There's a export type ToolUnion = Tool | ToolBash20250124 | ToolTextEditor20250124 | WebSearchTool20250305; from the Anthropic SDK

Tool is for custom tools, so it really widens the typing and doesn't seem like a great fit to assert if a Tool is one of the built in tools (see typing below).

Creating one and putting it in types.ts seems best (will put more details in your comment about the type property).

// From @anthropic-ai/sdk's messages.ts 
export interface Tool {
  /**
   * [JSON schema](https://json-schema.org/draft/2020-12) for this tool's input.
   *
   * This defines the shape of the `input` that your tool accepts and that the model
   * will produce.
   */
  input_schema: Tool.InputSchema;

  /**
   * Name of the tool.
   *
   * This is how the tool will be called by the model and in `tool_use` blocks.
   */
  name: string;

  /**
   * Create a cache control breakpoint at this content block.
   */
  cache_control?: CacheControlEphemeral | null;

  /**
   * Description of what this tool does.
   *
   * Tool descriptions should be as detailed as possible. The more information that
   * the model has about what the tool is and how to use it, the better it will
   * perform. You can use natural language descriptions to reinforce important
   * aspects of the tool input JSON schema.
   */
  description?: string;

  type?: 'custom' | null;
}

[benjamincburns]

I wonder if export type AnthropicBuiltInTool = Anthropic.Messages.Tool & { type : null } wouldn't be enough to resolve AnthropicBuiltInTool to what we'd be after, without having to explicitly redefine. We like to derive types (and schema) from the provider SDK when possible, as it makes forward compatibility a bit easier when they make additive changes.

[bleafman]

I pushed up a change that uses:

export type AnthropicBuiltInToolUnion = Exclude<
  Anthropic.Messages.ToolUnion,
  Anthropic.Messages.Tool
>;

This separates out the custom tool and should make it forward compatible since it's just deriving from the SDK types (theoretically Anthropic could put something unexpected in the union, but this at least the type inference to be "Everything besides the custom tools").

Let me know what you think about that one.

export type AnthropicBuiltInTool = Anthropic.Messages.Tool & { type : null }`` doesn't work because the built-in tools all have string literals with they timestamp (ex. text_editor_20250124) for the their type. The built-in tools and the custom tools don't actually overlap on the type` property.

type actually would let you narrow, but typing "Any string that isn't literal 'custom'" is pretty gross/kluggy 😓

[benjamincburns]

I don't see where built-in tools are enabled for this test. Were they meant to be configured in the initialization of chatModel above?

[benjamincburns]

To be clear, we don't want built-in tools to be enabled by default.

[bleafman]

Yeah that was an error on my end, built in tools are be enabled for the llm instance in all the tests now.

I'll also clean up the console.logs once I (hopefully) get the content blocks working like the OpenAI provider

[benjamincburns]

Have you seen how the built-in tool responses are formatted by the OpenAI provider when using the responses API? Would it be possible to make this structure look similar to that? This will cut down on the number of breaking changes we'll need to make when we eventually go to standardize this type of output.

[bleafman]

Ah! I didn't! That's helpful context. I'll see if I can get the structure to be similar.

[bleafman]

@benjamincburns I could use your confirmation/feedback before moving forward on how to split up / organize the data returned as part of the tool call.

You can see an example from Anthropic of a response with their built in web search tool here.

One of the big advantages of the built-in tool is that Claude will respond with text + citations side by side and you can use that render results (you're also required as part of the ToS to display the citations with any returned results).

Based on what's happening in the OpenAI provider, specifically in the _convertOpenAIResponseMessageToBaseMessage, it looks like we'd just append the citations as-is to an entry in the content array. Does that sound right?

It end up coming through like below:

 content: [
 {
      "citations": [
          {
              "type": "web_search_result_location",
              "cited_text": "Santolina, also known as cotton lavender, is a small, mounded shrub with silvery, aromatic foliage and button-like yellow flowers. It thrives in dry, ...",
              "url": "https://mygardeninspo.com/mediterranean-flowers/",
              "title": "26 Mediterranean Flowers",
              "encrypted_index": "..."
          }
      ],
      "type": "text",
      "text": "This is a small, mounded shrub with silvery, aromatic foliage and yellow button-like flowers. It's perfect for your needs as it thrives in dry, rocky soils and is often used for edging paths and borders. It prefers full sun and minimal watering"
  },
  // ...more entries
  ]

For the tool use itself, it seems like I should pass that into the tool_calls as a normal tool block, right?

I basically have the content working, the tool_calls are a little tricky since Anthropic expects a specific format in follow up turns but I have some ideas.

Just want to confirm before I go way down the wrong direction or I'm missing something.

Note: The OpenAI Provider does put some data into additional_kwargs.tool_calls but that field is marked as deprecated so I was planning on not replicating that behavior.

[benjamincburns]

Please also be sure to run yarn format && yarn lint && yarn build before committing.