Coding Agent (Azure Anthropic)

examples/coding-agent-azure-anthropic.rs shows the Azure-specific Anthropic provider path. It reads a full Azure Anthropic base URL or derives one from a resource name, targets the configured deployment name as the model, and keeps Anthropic features such as adaptive thinking, tool calls, retries, caching, and rate limiting.

Run

export AZURE_API_KEY="your-api-key"
export AZURE_ANTHROPIC_BASE_URL="https://your-resource.services.ai.azure.com/anthropic"
# or export AZURE_ANTHROPIC_RESOURCE="your-resource"
export AZURE_ANTHROPIC_AUTH_METHOD="x_api_key"
export AZURE_ANTHROPIC_MODEL="claude-opus-4-6"

cargo run --example coding-agent-azure-anthropic

Required and optional environment variables

• AZURE_ANTHROPIC_BASE_URL or AZURE_ANTHROPIC_RESOURCE is required.
• AZURE_ANTHROPIC_AUTH_METHOD defaults to x_api_key.
• AZURE_API_KEY is accepted as the practical fallback credential.
• AZURE_ANTHROPIC_API_KEY is used for explicit x-api-key auth.
• AZURE_ANTHROPIC_AUTH_TOKEN is used for explicit bearer-token auth.
• AZURE_ANTHROPIC_MODEL defaults to claude-opus-4-6.

The example constructs Azure Anthropic Messages API URLs in the form:

{base_url}/v1/messages

What this example actually configures

• LlmProvider::AzureAnthropic { base_url, auth_method }
• The deployment name from AZURE_ANTHROPIC_MODEL
• Anthropic adaptive thinking with EffortLevel::Max
• Prompt caching with a one-hour TTL
• Anthropic beta features for fine-grained tool streaming, context management, and interleaved thinking
• Anthropic tool choice, rate limiting, and retry configuration
• The shared coding tools: read_file, write_file, bash, and list_files

Key builder setup

let agent = AgentBuilder::new("azure-anthropic-coding-assistant")
    .provider(LlmProvider::AzureAnthropic {
        base_url: base_url.clone(),
        auth_method: auth_method.clone(),
    })
    .model(&model)
    .system_prompt(
        "You are an advanced coding assistant powered by Claude Opus 4.6 via Azure Anthropic. \
         You have access to file operations, bash commands, and directory listing. \
         Use your adaptive thinking capabilities to reason through complex problems. \
         Always explain your reasoning process and provide detailed analysis.",
    )
    .thinking(appam::llm::anthropic::ThinkingConfig::adaptive())
    .effort(appam::llm::anthropic::EffortLevel::Max)
    .caching(appam::llm::anthropic::CachingConfig {
        enabled: true,
        ttl: appam::llm::anthropic::CacheTTL::OneHour,
    })
    .with_tool(Arc::new(read_file()))
    .with_tool(Arc::new(write_file()))
    .with_tool(Arc::new(bash()))
    .with_tool(Arc::new(list_files()))
    .max_tokens(20000)
    .build()?;

Runtime behavior

• The binary fails fast if no Azure Anthropic endpoint can be resolved.
• It validates the credential shape required by the selected auth method before starting the session.
• The terminal UI prints reasoning, tool calls, and tool results as the stream progresses.