# Anthropic integration overview

> Generate text with Anthropic's Claude models inside a Persona Workflow, passing case data into prompts for summaries, analysis, and decision support.

Source: https://help.withpersona.com/articles/4ItMteIsKhrKaEyBIg76l8/
Section: Marketplace and 3rd-Party Integrations > AI > Anthropic

## Overview

[Anthropic](https://www.anthropic.com/) provides access to **Claude**, a family of large language models (LLMs) designed for text and content generation. Persona’s integration with Anthropic allows teams to generate AI-powered responses directly within workflows using the **Generate Text** action.

This integration enables real-time summarization, decision support, and automation, which helps teams work faster and more consistently across cases.

## Benefits

**Automated Text Generation:** Automatically create summaries, responses, or decisions based on workflow context and inputs.

**Operational Efficiency:** Reduce manual review time by letting Claude handle routine text drafting or analysis.

**Context-Aware Responses:** Pass case or account details into prompts for highly personalized and accurate results.

## Integration Features

Persona’s Anthropic integration supports intelligent text generation through Anthropic’s **Generate Text** endpoint. Key capabilities include:

- **Generate Text:** Send prompts, messages, or media to Claude and receive text-based responses in real time.
- **Dynamic Field Mapping:** Insert Persona data (like notes, explanations, or tags) into prompts for contextual output.
- **Model Flexibility:** Select from the latest Claude model variants for your use case.
- **Customizable Parameters:** Control temperature, token limit, and sampling behavior to balance creativity and precision.

## Setting up the Anthropic Integration

### Prerequisites

Before setting up, ensure you have:

- An **Anthropic API Key** (found in your Anthropic account)
- Admin access to the **Persona Dashboard**
- Permission to edit and publish workflows

### Setting up Anthropic Credentials

1.  In the Persona Dashboard, navigate to **Integrations > Marketplace** and click on **Anthropic**.
2.  Click **Add Credential** at the top of the page.
3.  Enter a **nickname** for your credential.
4.  Paste your **Anthropic API key** when prompted.
5.  Click **Continue** to authenticate and connect.

Once connected, Anthropic will appear in your list of available workflow integrations.

## Using the Anthropic Integration in a Workflow

1.  Open an existing workflow or create a new one.
2.  Add a new **Action Step > Integrations > Anthropic**.
3.  Select the **Generate Text** operation.
4.  Fill in the configuration fields outlined below.
5.  Save and publish your workflow.

When triggered, the workflow will send your input to Anthropic and return the generated text automatically.

## Anthropic Operations Overview

### Generate Text

Generates a text-based response from Anthropic’s Claude model using text and/or image inputs. Useful for automating reasoning, summarization, or analysis steps within a workflow.

### Configuration Steps

Provide values for the required fields:

- **Model Variant ID**– Specify which Claude model variant to use. View available options in [Anthropic’s model documentation](https://docs.anthropic.com/).
- **Max Output Tokens**– Set the maximum number of tokens the model can return.
- **Messages**– Define the conversation context that the model should consider.

Each message includes:

- **Role**– Choose whether the message is from the `user` or `assistant`.
- **Content**– Specify whether you’re providing text or an image.

If **content type = image**, provide:

- **Source Type**– Only `url` is available.
- **Image URL**– Paste the image link you’d like Claude to process.

If **content type = text**, provide:

- **Text**– The written input or prompt you want the model to respond to.

Optionally, configure additional parameters:

- **Temperature** – Adjusts creativity/randomness of the output.
- **Top P** – Controls diversity of generated tokens.
- **Top K** – Sets the number of possible tokens considered per step.

## FAQs

**Can I send both text and images in a single request?**

Yes, the Generate Text endpoint supports multi-modal inputs, so you can pass both text and image content in your messages for Claude to reference in its response. Just add an additional context.

**What happens if my output exceeds the token limit?**

If the model output exceeds your configured Max Output Tokens, the response will be truncated. You can increase the limit (up to Anthropic’s maximum) to allow longer responses.

## Plans Explained

### Anthropic Integration by plan

|                       | Startup Program | Essential Plan | Growth Plan | Enterprise Plan |
| --------------------- | --------------- | -------------- | ----------- | --------------- |
| Anthropic Integration | Not Available   | Not Available  | Limited     | Available       |

[Learn more about pricing and plans](https://withpersona.com/pricing?utm_source=product&utm_medium=referral&utm_audience=a&utm_campaign=cm_gen_ds_hc-plan-table).