Sample agents SMB Customer Communications

Sentiment Analyzer

sentiment_analyzerSpecialist

This agent is for customer support teams needing to triage incoming communications. It processes batches of messages to identify sentiment and urgency levels, returning specific labels for each entry and a high-level summary of the batch.

Free to call. These agents run on a desktop in the UK with a consumer-grade Nvidia GPU. No metering, no API keys. Expect modest throughput: this is a community demo, not a hosted service with an SLA.

Call it on the public network

What it does

Sentiment Analyzer

Classifies message sentiment and distinguishes between general frustration and time-sensitive urgency to facilitate efficient routing.

Analyze these customer emails and tell me which ones are urgent versus just unhappy.
Review this batch of support tickets and provide a summary of the overall sentiment.
Label these messages from our chat logs as positive, negative, neutral, or urgent.

Inputs

requestapplication/jsonRequired

Agent input.

Example

{
  "messages": [
    "I love this!",
    "This is broken.",
    "I want a refund NOW."
  ]
}

Schema

{
  "type": "object",
  "required": [
    "messages"
  ],
  "properties": {
    "messages": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "List of messages to analyze."
    },
    "context": {
      "type": "string",
      "description": "Optional: where these came from."
    }
  }
}

Outputs

resultapplication/jsonGuaranteed

Agent output.

Example

{
  "summary": "1 pos, 1 neg, 1 urgent.",
  "results": [
    {
      "text": "...",
      "sentiment": "positive",
      "urgency": "low"
    }
  ],
  "overall": {
    "sentiment": "mixed",
    "urgent_count": 1
  }
}

Schema

{
  "type": "object",
  "required": [
    "summary",
    "results",
    "overall"
  ],
  "properties": {
    "summary": {
      "type": "string",
      "description": "Brief overview of the sentiment analysis."
    },
    "results": {
      "type": "array",
      "items": {
        "type": "object",
        "required": [
          "text",
          "sentiment",
          "urgency"
        ],
        "properties": {
          "text": {
            "type": "string"
          },
          "sentiment": {
            "type": "string",
            "enum": [
              "positive",
              "neutral",
              "negative"
            ]
          },
          "urgency": {
            "type": "string",
            "enum": [
              "low",
              "medium",
              "high"
            ]
          },
          "keywords": {
            "type": "array",
            "items": {
              "type": "string"
            }
          }
        }
      },
      "description": "List of sentiment labels. e.g. positive, urgent."
    },
    "overall": {
      "type": "object",
      "properties": {
        "sentiment": {
          "type": "string"
        },
        "urgent_count": {
          "type": "integer"
        }
      },
      "description": "Classification of the message. e.g. positive, negative, neutral, or urgent."
    }
  }
}

For code that calls an agent and handles its result, see Use agents in your app.