Anthropic Chat Models
Anthropic is an AI safety and research company. It provides the Claude family of Large Language Models, designed with constitutional AI principles for safe and controllable output.
This extension allows you to integrate Claude models into your Quarkus applications via the Anthropic API.
Prerequisites
To use Anthropic models, you need an API key. Follow the steps on the Claude documentation portal to request access and retrieve your credentials.
Dependency
To enable Anthropic LLM integration in your project, add the following dependency:
<dependency>
<groupId>io.quarkiverse.langchain4j</groupId>
<artifactId>quarkus-langchain4j-anthropic</artifactId>
<version>1.8.4</version>
</dependency>Even better, if you use the Quarkus platformn BOM (default for projects generated), add the Quarkus Langchain4J BOM and all dependency versions will align:
<dependencyManagement>
<dependencies>
<dependency>
<groupId>${quarkus.platform.group-id}</groupId>
<artifactId>${quarkus.platform.artifact-id}</artifactId>
<version>${quarkus.platform.version}</version>
<type>pom</type>
<scope>import</scope>
</dependency>
<dependency>
<groupId>${quarkus.platform.group-id}</groupId>
<artifactId>quarkus-langchain4j-bom</artifactId> (1)
<version>${quarkus.platform.version}</version> (2)
<type>pom</type>
<scope>import</scope>
</dependency>
</dependencies>
</dependencyManagement>
<dependencies>
<dependency>
<groupId>io.quarkiverse.langchain4j</groupId>
<artifactId>quarkus-langchain4j-anthropic</artifactId>
(3)
</dependency>
</dependencies>| 1 | In your dependencyManagement section, add the quarkus-langchain4j-bom |
| 2 | Inherit the version from your platform version |
| 3 | Voilà, no need for version alignment anymore |
If no other LLM extension is installed, AI Services will automatically use the configured Anthropic chat model.
Configuration
Set your API key in the application.properties file:
quarkus.langchain4j.anthropic.api-key=...You can also set it using the environment variable:
QUARKUS_LANGCHAIN4J_ANTHROPIC_API_KEY=...By default, the extension uses the latest available Claude model. You can specify the model explicitly using:
quarkus.langchain4j.anthropic.chat-model.model-name=claude-opus-4-20250514Refer to Anthropic’s model catalog for available versions, such as:
-
claude-sonnet-4-20250514 -
claude-3-opus-20240229 -
claude-3-haiku-20240307
Usage
You can inject the chat model directly:
@Inject ChatModel chatModel;Or declare an AI service interface:
@RegisterAiService
public interface Assistant {
String chat(String input);
}Advanced Tool Use Features
These features enable more efficient tool orchestration, improved accuracy, and reduced token consumption.
All features are optional and can be enabled independently.
Tool Search
Anthropic’s Tool Search Tool allows Claude to use search tools to access thousands of tools without consuming its context window
The Tool Search Tool lets Claude dynamically discover tools instead of loading all definitions upfront. You provide all your tool definitions to the API, but mark tools with defer_loading: true to make them discoverable on-demand. Deferred tools aren’t loaded into Claude’s context initially. Claude only sees the Tool Search Tool itself plus any tools with defer_loading: false (your most critical, frequently-used tools).
When Claude needs specific capabilities, it searches for relevant tools. The Tool Search Tool returns references to matching tools, which get expanded into full definitions in Claude’s context.
To enable this feature, set the following property:
quarkus.langchain4j.anthropic.chat-model.tool-search.enabled=trueYou can optionally specify the search type (regex is the default, or bm25):
quarkus.langchain4j.anthropic.chat-model.tool-search.type=bm25|
Under the hood, when Tool Search is enabled, the Quarkus extension automatically:
|
When tool search is enabled, you can mark specific tools to be discovered on-demand using the defer_loading metadata:
@Tool(name = "get_weather", value = "Get the weather for a city", metadata = "{\"defer_loading\": true}")
public String getWeather(String city) {
// ...
}Programmatic Tool Calling
Anthropic’s Programmatic Tool Calling allows Claude to invoke tools in a code execution environment reducing the impact on the model’s context window.
Programmatic Tool Calling enables Claude to orchestrate tools through code rather than through individual API round-trips. Instead of Claude requesting tools one at a time with each result being returned to its context, Claude writes code that calls multiple tools, processes their outputs, and controls what information actually enters its context window.
Claude excels at writing code and by letting it express orchestration logic in Python rather than through natural language tool invocations, you get more reliable, precise control flow. Loops, conditionals, data transformations, and error handling are all explicit in code rather than implicit in Claude’s reasoning.
To enable this feature, set the following property:
quarkus.langchain4j.anthropic.chat-model.programmatic-tool-calling.enabled=true|
Under the hood, when Programmatic Tool Calling is enabled, the extension automatically:
|
Tools that should be callable from generated code must include the
allowed_callers metadata:
@Tool(
name = "get_weather",
value = "Get the weather for a city",
metadata = "{\"allowed_callers\": [\"code_execution_20250825\"]}"
)
public String getWeather(String city) {
// ...
}Tool Use Examples
This feature provides a universal standard for demonstrating how to effectively use a given tool.
Tool Use Examples let you provide sample tool calls directly in your tool definitions. Instead of relying on schema alone, you show Claude concrete usage patterns.
For example, format ambiguity:
if a date is passed as string, should it use "2024-11-06", "Nov 6, 2024", or "2024-11-06T00:00:00Z"?
To enable this feature, set the following property:
quarkus.langchain4j.anthropic.chat-model.tool-use-examples.enabled=true|
Under the hood, when Tool Use Examples is enabled, the extension automatically:
|
Tools that include examples must define the input_examples metadata:
private static final String METADATA = """
{
"input_examples": [
{
"title": "Login page returns 500 error",
"priority": "critical",
"labels": ["bug", "authentication"],
"date": "2024-11-06"
},
{
"title": "Update API documentation",
"date": "2024-12-01"
},
{
"title": "Brainstorming session"
}
]
}
""";
@Tool(name = "create_ticket", value = "Create a support ticket", metadata = METADATA)
public String createTicket(String title, String priority, String date) {
return "TICKET-123";
}Configuration Reference
Configuration property fixed at build time - All other configuration properties are overridable at runtime
Configuration property |
Type |
Default |
|---|---|---|
Whether the model should be enabled Environment variable: |
boolean |
|
Base URL of the Anthropic API Environment variable: |
string |
|
Anthropic API key Environment variable: |
string |
|
The Anthropic version Environment variable: |
string |
|
If set to Environment variable: |
boolean |
|
Timeout for Anthropic calls Environment variable: |
|
|
Whether the Anthropic client should log requests Environment variable: |
boolean |
|
Whether the Anthropic client should log responses Environment variable: |
boolean |
|
Whether the Anthropic client should log requests as cURL commands Environment variable: |
boolean |
|
Whether to enable the integration. Defaults to Environment variable: |
boolean |
|
Model name to use Environment variable: |
string |
|
What sampling temperature to use, between 0.0 and 1.0. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. It is generally recommended to set this or the Environment variable: |
double |
|
The maximum number of tokens to generate in the completion. The token count of your prompt plus Environment variable: |
int |
|
Double (0.0-1.0). Nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. It is generally recommended to set this or the Environment variable: |
double |
|
Reduces the probability of generating nonsense. A higher value (e.g. 100) will give more diverse answers, while a lower value (e.g. 10) will be more conservative Environment variable: |
int |
|
The maximum number of times to retry. 1 means exactly one attempt, with retrying disabled. Environment variable: |
int |
|
The custom text sequences that will cause the model to stop generating Environment variable: |
list of string |
|
Whether chat model requests should be logged Environment variable: |
boolean |
|
Whether chat model responses should be logged Environment variable: |
boolean |
|
Cache system messages to reduce costs for repeated prompts. Requires minimum 1024 tokens (Claude Opus/Sonnet) or 2048-4096 tokens (Haiku). Supported models: Claude Opus 4.1, Sonnet 4.5, Haiku 4.5, and later models. Environment variable: |
boolean |
|
Cache tool definitions to reduce costs. Requires minimum 1024 tokens (Claude Opus/Sonnet) or 2048-4096 tokens (Haiku). Supported models: Claude Opus 4.1, Sonnet 4.5, Haiku 4.5, and later models. Environment variable: |
boolean |
|
The thinking type to enable Claude’s reasoning process Environment variable: |
string |
|
The token budget for the model’s thinking process Environment variable: |
int |
|
Whether thinking results should be returned in the response Environment variable: |
boolean |
|
Whether previously stored thinking should be sent in follow-up requests Environment variable: |
boolean |
|
Enable interleaved thinking for Claude 4 models, allowing reasoning between tool calls. Requires Claude 4 model (e.g., claude-opus-4-20250514) and thinking.type: enabled. Environment variable: |
boolean |
|
Enable Anthropic’s Tool Search Tool for on-demand tool discovery. When enabled, this automatically adds the tool search server tool, sets the required beta header, and enables the "defer_loading" tool metadata key. Tools annotated with `@Tool(metadata = "{\"defer_loading\": true`")} will be discovered on demand instead of loaded upfront. Environment variable: |
boolean |
|
The type of tool search to use. Available types: "regex" (default) or "bm25". Environment variable: |
string |
|
Enable Anthropic’s Programmatic Tool Calling via the Code Execution server tool. When enabled, this automatically adds the code execution server tool, the Tools that should be callable from code must include: `@Tool(metadata = "{\"allowed_callers\": [\"code_execution_20250825\"]`")} Environment variable: |
boolean |
|
Enable Anthropic’s Tool Use Examples feature. When enabled, the Tools with examples must include: Environment variable: |
boolean |
|
Type |
Default |
|
Base URL of the Anthropic API Environment variable: |
string |
|
Anthropic API key Environment variable: |
string |
|
The Anthropic version Environment variable: |
string |
|
If set to Environment variable: |
boolean |
|
Timeout for Anthropic calls Environment variable: |
|
|
Whether the Anthropic client should log requests Environment variable: |
boolean |
|
Whether the Anthropic client should log responses Environment variable: |
boolean |
|
Whether the Anthropic client should log requests as cURL commands Environment variable: |
boolean |
|
Whether to enable the integration. Defaults to Environment variable: |
boolean |
|
Model name to use Environment variable: |
string |
|
What sampling temperature to use, between 0.0 and 1.0. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. It is generally recommended to set this or the Environment variable: |
double |
|
The maximum number of tokens to generate in the completion. The token count of your prompt plus Environment variable: |
int |
|
Double (0.0-1.0). Nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. It is generally recommended to set this or the Environment variable: |
double |
|
Reduces the probability of generating nonsense. A higher value (e.g. 100) will give more diverse answers, while a lower value (e.g. 10) will be more conservative Environment variable: |
int |
|
The maximum number of times to retry. 1 means exactly one attempt, with retrying disabled. Environment variable: |
int |
|
The custom text sequences that will cause the model to stop generating Environment variable: |
list of string |
|
Whether chat model requests should be logged Environment variable: |
boolean |
|
Whether chat model responses should be logged Environment variable: |
boolean |
|
Cache system messages to reduce costs for repeated prompts. Requires minimum 1024 tokens (Claude Opus/Sonnet) or 2048-4096 tokens (Haiku). Supported models: Claude Opus 4.1, Sonnet 4.5, Haiku 4.5, and later models. Environment variable: |
boolean |
|
Cache tool definitions to reduce costs. Requires minimum 1024 tokens (Claude Opus/Sonnet) or 2048-4096 tokens (Haiku). Supported models: Claude Opus 4.1, Sonnet 4.5, Haiku 4.5, and later models. Environment variable: |
boolean |
|
The thinking type to enable Claude’s reasoning process Environment variable: |
string |
|
The token budget for the model’s thinking process Environment variable: |
int |
|
Whether thinking results should be returned in the response Environment variable: |
boolean |
|
Whether previously stored thinking should be sent in follow-up requests Environment variable: |
boolean |
|
Enable interleaved thinking for Claude 4 models, allowing reasoning between tool calls. Requires Claude 4 model (e.g., claude-opus-4-20250514) and thinking.type: enabled. Environment variable: |
boolean |
|
Enable Anthropic’s Tool Search Tool for on-demand tool discovery. When enabled, this automatically adds the tool search server tool, sets the required beta header, and enables the "defer_loading" tool metadata key. Tools annotated with `@Tool(metadata = "{\"defer_loading\": true`")} will be discovered on demand instead of loaded upfront. Environment variable: |
boolean |
|
The type of tool search to use. Available types: "regex" (default) or "bm25". Environment variable: |
string |
|
Enable Anthropic’s Programmatic Tool Calling via the Code Execution server tool. When enabled, this automatically adds the code execution server tool, the Tools that should be callable from code must include: `@Tool(metadata = "{\"allowed_callers\": [\"code_execution_20250825\"]`")} Environment variable: |
boolean |
|
Enable Anthropic’s Tool Use Examples feature. When enabled, the Tools with examples must include: Environment variable: |
boolean |
|
|
About the Duration format
To write duration values, use the standard You can also use a simplified format, starting with a number:
In other cases, the simplified format is translated to the
|