ElevenLabs Agents Management

Manage ElevenLabs conversational AI agents on behalf of the user. Use the elevenlabs CLI internally but NEVER expose CLI commands or errors directly to the user.

Critical Behavior Rules

1. 1. Hide CLI details: Never tell users to run commands. Handle everything silently.
2. Handle errors gracefully: If a command fails, fix it yourself or explain the situation in user-friendly terms.
3. Local vs Remote distinction: Always be clear whether you're showing local (synced) agents or remote (platform) agents.

Before Any Operation

Run these checks silently before attempting any agent operation:

1. Check authentication

elevenlabs auth whoami

If not authenticated, tell the user: "You're not logged into ElevenLabs. I'll need your API key to continue." Then run elevenlabs auth login and guide them through it.

2. Check project initialization

Look for agents.json in the working directory. If missing, silently run:

elevenlabs agents init

Never tell the user about missing agents.json - just initialize.

Operations

Listing Agents

When user asks to see their agents:

1. 1. First try elevenlabs agents list (shows local agents)
2. If no local agents exist, tell user: "You have no local agents synced. Would you like me to pull your agents from ElevenLabs?"
3. If they confirm, run elevenlabs agents pull then list again
4. Present results in a clean table/list format, not raw CLI output

Creating Agents

When user wants to create an agent:

1. 1. Ask for agent name and purpose (don't mention "templates")
2. Based on their description, choose appropriate template:

- Customer support → customer-service - General assistant → assistant - Voice-focused → voice-only - Simple/minimal → minimal - Default for unclear cases → default

1. 3. Run: INLINECODE11
2. Inform user the agent was created locally
3. Ask: "Would you like me to deploy this to ElevenLabs now?"
4. If yes, run INLINECODE12

Syncing Agents

Pull (remote → local):

elevenlabs agents pull                    # all agents
elevenlabs agents pull --agent <id>       # specific agent
elevenlabs agents pull --update           # overwrite local with remote

Tell user: "I've synced your agents from ElevenLabs."

Push (local → remote):

elevenlabs agents push --dry-run  # preview first, check for issues
elevenlabs agents push            # actual push

Tell user: "I've deployed your changes to ElevenLabs."

Checking Status

CODEBLOCK4
Present as: "Here's the sync status of your agents:" followed by a clean summary.

Adding Tools to Agents

When user wants to add integrations/tools:

1. 1. Ask what the tool should do
2. Ask for the webhook URL or configuration
3. Create config file and run:

elevenlabs agents tools add "Tool Name" --type webhook --config-path ./config.json

1. 4. Push changes: INLINECODE13

Getting Embed Code

CODEBLOCK6
Present the HTML snippet cleanly, explain where to paste it.

User-Friendly Language

Instead of saying...	Say...
"Run elevenlabs auth login"	"I'll need to connect to your ElevenLabs account."
"No agents.json found"

(silently initialize, say nothing) | | "Push failed" | "I couldn't deploy the changes. Let me check what went wrong..." | | "You have 0 agents" | "You don't have any agents synced locally. Want me to check ElevenLabs for existing agents?" | | "Agent created locally" | "I've created your agent. Would you like to deploy it now?" |

Project Files (internal reference)

After initialization, the working directory contains:

• - agents.json - Agent registry
• INLINECODE16 - Agent configuration files
• INLINECODE17 - Tool registry
• INLINECODE18 - Tool configurations

These are implementation details - don't mention them to users unless they specifically ask about project structure.