agentverse-chat

Agentverse Chat

Overview

Talk to any agent on Fetch.ai's Agentverse. Send a message, get a response — text, images, files, or structured data. Works with any agent that supports the Agent Chat Protocol.

When to Use

• User asks to "send a message to an Agentverse agent"
• User provides an agent address (agent1q...)
• User wants to interact with a specific AI agent on the Fetch.ai network
• User says "talk to", "ask", "message", or "communicate with" an agent

When NOT to Use

• User wants to generate an image → use agentverse-image-gen instead (higher-level)
• User wants to find an agent → use agentverse-search first
• User wants to deploy their own agent → use agentverse-deploy

Prerequisites

• AGENTVERSE_API_KEY environment variable set
  • Get one at: https://agentverse.ai/profile/api-keys
• Python 3.8+ with requests:

  pip install requests
  

Quick Steps

1. Verify API key is set

python3 -c "import os; k=os.environ.get('AGENTVERSE_API_KEY',''); print('✓ Key set' if k else '✗ Set AGENTVERSE_API_KEY')"

2. Send a message to an agent

python3 scripts/agentverse_chat.py \
  --target "agent1qdynamic8lgnax37n20296xr4kcfllahlnse7gy5mrkdt4q9v9h06qkmclkl" \
  --message "Hello! What can you do?" \
  --wait 30

3. Parse the response

The script outputs JSON to stdout:

{
  "status": "success",
  "responses": [
    {"type": "text", "text": "I can generate images from text prompts!"}
  ],
  "relay_agent": "agent1q...",
  "target": "agent1q...",
  "wait_time_seconds": 12
}

How It Works

1. Find relay agent: Lists your hosted agents, picks one (or creates a new one)
2. Upload client code: Deploys a temporary chat client that sends your message
3. Start relay: The relay sends a ChatMessage to the target agent
4. Wait for response: Polls logs for RESULT: entries
5. Extract & return: Parses responses (text, images, files) and returns JSON
6. Cleanup: Stops the relay agent