Skip to main content

Anthropic Claude Setup

AxonFlow supports Anthropic's Claude models for LLM routing and orchestration. Claude is available in the Community edition and is known for its safety-focused design and long context capabilities.

Prerequisites​

Quick Start​

1. Get API Key​

  1. Go to Anthropic Console
  2. Click "Create Key"
  3. Copy the generated key

2. Configure AxonFlow​

# Required
export ANTHROPIC_API_KEY=sk-ant-your-api-key-here

# Optional: Specify model (default: claude-3-5-sonnet-20241022)
export ANTHROPIC_MODEL=claude-3-5-haiku-20241022

3. Start AxonFlow​

docker compose up -d

Supported Models​

ModelContext WindowBest For
claude-sonnet-4-20250514200K tokensLatest, balanced quality/speed
claude-opus-4-20250514200K tokensHighest capability
claude-3-5-sonnet-20241022200K tokensFast, high quality (default)
claude-3-5-haiku-20241022200K tokensFastest, cost-effective
claude-3-opus-20240229200K tokensComplex reasoning
claude-3-sonnet-20240229200K tokensBalanced (legacy)
claude-3-haiku-20240307200K tokensFast responses (legacy)

Configuration Options​

Environment Variables​

VariableRequiredDefaultDescription
ANTHROPIC_API_KEYYes-Anthropic API key
ANTHROPIC_MODELNoclaude-3-5-sonnet-20241022Default model
ANTHROPIC_ENDPOINTNohttps://api.anthropic.comAPI endpoint
ANTHROPIC_TIMEOUT_SECONDSNo120Request timeout (seconds)
Model Configuration

Always set ANTHROPIC_MODEL explicitly if your API key has limited model access (e.g., only claude-3-haiku-20240307). AxonFlow uses the configured model for all requests to this provider.

YAML Configuration​

For more control, use YAML configuration:

# axonflow.yaml
llm_providers:
  anthropic:
    enabled: true
    config:
      model: claude-3-5-sonnet-20241022
      max_tokens: 8192
      timeout: 120s
    credentials:
      api_key: ${ANTHROPIC_API_KEY}
    priority: 10
    weight: 0.5

Capabilities​

The Anthropic provider supports:

  • Chat completions - Conversational AI
  • Streaming responses - Real-time token streaming
  • Long context - Up to 200K tokens
  • Vision - Image understanding
  • Tool use - Function calling
  • Code generation - Programming assistance
  • Constitutional AI - Built-in safety alignment

Usage Examples​

Proxy Mode (Python SDK)​

Proxy mode routes requests through AxonFlow for simple integration:

from axonflow import AxonFlow

async with AxonFlow(agent_url="http://localhost:8080") as client:
    response = await client.execute_query(
        user_token="user-123",
        query="Explain quantum computing",
        request_type="chat",
        context={"provider": "anthropic", "model": "claude-3-5-sonnet-20241022"}
    )
    print(response.content)

Proxy Mode (cURL)​

curl -X POST http://localhost:8080/api/request \
  -H "Content-Type: application/json" \
  -H "X-User-Token: user-123" \
  -d '{
    "query": "What is quantum computing?",
    "provider": "anthropic",
    "model": "claude-3-5-sonnet-20241022",
    "max_tokens": 500
  }'

Gateway Mode (TypeScript SDK)​

Gateway mode gives you full control over the LLM call while AxonFlow handles policy enforcement and audit logging:

import { AxonFlow } from '@axonflow/sdk';
import Anthropic from '@anthropic-ai/sdk';

const axonflow = new AxonFlow({
  endpoint: 'http://localhost:8080',
  apiKey: 'your-axonflow-key'
});

// 1. Pre-check: Get policy approval
const ctx = await axonflow.getPolicyApprovedContext({
  userToken: 'user-123',
  query: 'Explain quantum computing'
});

if (!ctx.approved) {
  throw new Error(`Blocked: ${ctx.blockReason}`);
}

// 2. Call Anthropic directly
const anthropic = new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY });
const message = await anthropic.messages.create({
  model: 'claude-3-5-sonnet-20241022',
  max_tokens: 1024,
  messages: [{ role: 'user', content: ctx.approvedData.query }]
});
const response = message.content[0].text;

// 3. Audit the call
await axonflow.auditLLMCall({
  contextId: ctx.contextId,
  responseSummary: response.substring(0, 100),
  provider: 'anthropic',
  model: 'claude-3-5-sonnet-20241022',
  tokenUsage: {
    promptTokens: message.usage.input_tokens,
    completionTokens: message.usage.output_tokens,
    totalTokens: message.usage.input_tokens + message.usage.output_tokens
  },
  latencyMs: 250
});

Streaming​

Anthropic supports server-sent events (SSE) for streaming responses:

import Anthropic from '@anthropic-ai/sdk';

const anthropic = new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY });

const stream = await anthropic.messages.stream({
  model: 'claude-3-5-sonnet-20241022',
  max_tokens: 1024,
  messages: [{ role: 'user', content: 'Write a story' }]
});

for await (const chunk of stream) {
  if (chunk.type === 'content_block_delta') {
    process.stdout.write(chunk.delta.text);
  }
}

Pricing​

Anthropic pricing (as of December 2025):

ModelInput (per 1M tokens)Output (per 1M tokens)
Claude Sonnet 4$3.00$15.00
Claude Opus 4$15.00$75.00
Claude 3.5 Sonnet$3.00$15.00
Claude 3.5 Haiku$0.80$4.00
Claude 3 Opus$15.00$75.00

AxonFlow provides cost estimation via the /api/cost/estimate endpoint.

Multi-Provider Routing​

Configure Anthropic alongside other providers for intelligent routing:

llm_providers:
  anthropic:
    enabled: true
    config:
      model: claude-3-5-sonnet-20241022
    credentials:
      api_key: ${ANTHROPIC_API_KEY}
    priority: 100

  openai:
    enabled: true
    config:
      model: gpt-4o
    credentials:
      api_key: ${OPENAI_API_KEY}
    priority: 50

routing:
  strategy: priority
  fallback_enabled: true

Health Checks​

Check the Anthropic provider health status:

# Check specific provider health
curl http://localhost:8081/api/v1/llm-providers/anthropic/health

Response:

{
  "provider": "anthropic",
  "status": "healthy",
  "latency_ms": 52,
  "model": "claude-3-5-sonnet-20241022"
}

To check all configured providers at once:

curl http://localhost:8081/api/v1/llm-providers/status

Response:

{
  "providers": {
    "anthropic": {
      "status": "healthy",
      "latency_ms": 52,
      "model": "claude-3-5-sonnet-20241022"
    },
    "openai": {
      "status": "healthy",
      "latency_ms": 45,
      "model": "gpt-4o"
    }
  }
}

Error Handling​

Common error codes from Anthropic:

StatusReasonAction
400Invalid requestCheck request format
401Invalid API keyVerify ANTHROPIC_API_KEY
429Rate limit exceededImplement backoff/retry
500Server errorRetry with exponential backoff
529API overloadedRetry with backoff

AxonFlow automatically handles retries for transient errors (429, 500, 529).

Best Practices​

  1. Use appropriate models - Sonnet for most tasks, Haiku for speed, Opus for complex reasoning
  2. Set reasonable timeouts - 120s default is good for most use cases
  3. Enable fallback providers - Configure OpenAI/Gemini as backup
  4. Monitor costs - Use AxonFlow's cost dashboard to track usage
  5. Leverage long context - Claude handles up to 200K tokens well

Troubleshooting​

"Invalid API key"​

  • Verify the key at Anthropic Console
  • Ensure the key hasn't been disabled
  • Check for leading/trailing whitespace

"Model not found"​

  • Verify model name format (e.g., claude-3-5-sonnet-20241022)
  • Check if model is available in your region
  • Note: Older model versions may be deprecated

"Rate limit exceeded"​

  • Check usage at Anthropic Console
  • Request a rate limit increase
  • Implement exponential backoff

Next Steps​