Docs / Strand / connectors/ai-anthropic

Anthropic Claude Connector

Send prompts to Anthropic Claude models and receive completions from your Strand workflows. Use this connector for analysis, writing, code generation, classification, and any task that benefits from a large language model.

Direction: Read / Write | Type: ai.anthropic

Prerequisites

You need an Anthropic API key:

1. Go to console.anthropic.com/settings/keys
2. Click Create Key
3. Copy the key (it starts with sk-ant-)

Connector Setup

Create an Anthropic Claude connector from the Connectors page.

Configuration Fields

Note: Anthropic's temperature range is 0--1, narrower than OpenAI and Google (0--2).

Operations

Chat Completion

Generate a chat completion response from a Claude model. This is the single operation exposed by the connector.

Operation ID: chat

The prompt is sent as a user message. If System Instructions are configured on the node, they are sent as the system message preceding the prompt.

Models

You can use any valid Anthropic model name -- the field is free text, so newer models work as soon as Anthropic releases them, and older ones keep working until Anthropic retires them. Use the plain model ID without a date suffix.

Node Configuration

When you drag an Anthropic Claude connector node onto the canvas, these fields are available in the node inspector. They apply to this specific node and override connector-level defaults where noted.

Output

A successful chat completion returns:

{
  "response": "Based on the data provided, there are three key findings...",
  "model": "claude-sonnet-5",
  "provider": "anthropic",
  "usage": {
    "prompt_tokens": 150,
    "completion_tokens": 89,
    "total_tokens": 239
  }
}

When memory is enabled, the output includes "memory_enabled": true. When MCP tools are called, the output includes an "mcp_tools_called" array. If retries occurred, a "retries" count is included.

Access the response in downstream nodes:

{{ steps.NODE_ID.output_payload.response }}
{{ steps.NODE_ID.output_payload.usage.total_tokens }}

Example Workflow

Summarize daily sensor data with ongoing context:

1. Schedule Trigger fires every evening at 6 PM
2. HTTP node fetches the day's sensor readings from your API
3. Anthropic Claude node analyzes the data:

• System Instructions: You are an IoT data analyst. Summarize sensor readings and flag anomalies. When memory is on, reference trends from previous days.
• Prompt: Today's sensor data:\n{{ payload.readings | tojson }}
• Temperature: 0.3
• Memory Enabled: true
• Memory TTL: 604800 (7 days)
• Memory Max Messages: 30

1. Slack node posts the summary to #sensor-reports
2. Router node checks for flagged anomalies and sends urgent alerts if found

Because memory is enabled, the model builds on its prior analysis each day and can identify trends over the week.

Retries

The node automatically retries on transient errors -- rate limits (429), server errors (5xx), timeouts, and connection failures. Retries use exponential backoff with jitter and respect Retry-After headers from Anthropic. Non-retryable errors (invalid API key, bad request) fail immediately.

Conversation Memory

Toggle Memory Enabled on and the node will remember every exchange across workflow runs. Each node gets its own isolated, account-scoped memory that is securely persisted across runs. See AI Connectors for full details on memory behavior, TTL, trimming, and when to enable it.

MCP Tool Access

AI nodes can call external tools via MCP servers. Add one or more MCP server connectors to the MCP Servers field and the model will have access to those tools during its response.

Errors

Related

• AI Connectors -- shared node config, memory, MCP, and examples for all providers
• OpenAI | Google Gemini
• MCP Server -- for tool access

← Back to Strand docs