ElevenLabs

Overview

Test your ElevenLabs agents using multiple methods: phone numbers, websockets, or chat interfaces. Choose the testing method that best fits your use case.

Configure automatically

Select ElevenLabs

In Cekura, go to Agents, create an agent, and select ElevenLabs as the provider.

Enter your ElevenLabs credentials

Enter the ElevenLabs Agent ID and an API key with access to Agents. Select US for the default API or EU Residency if your ElevenLabs workspace uses the EU endpoint.

Automatically configure an ElevenLabs agent and select its API region in Cekura

Configure the agent

Click Configure Agent. Cekura imports the available name, prompt, language, first message, phone number, tools, dynamic variables, and knowledge base.

Review the connection

When the import finishes, open the agent and confirm the imported settings before running an evaluator.

To enter the Agent ID and API key yourself, click Set up manually instead and follow the connection-specific steps below.

Phone Number
Websocket
Chat
Code

Test your agents using phone numbers directly from the frontend.

Configure phone number

Go to your agent settings and configure the phone number for testing:

Enter the phone number associated with your ElevenLabs agent. No additional credentials required for phone-based testing.

Run tests from frontend

Select scenarios and run tests using the UI. This will open the Configure Run dialog.In the Configure Run dialog, select Telephony under Voice connections and click Run.

Cekura automatically:

Initiates phone calls to your agent
Executes test scenarios
Captures audio and transcript data

You can override the phone number directly in this dialog before running.

Connection options that appear greyed out have not been configured in Agent Settings yet — set them up there to enable them.

View results

Results appear in your dashboard. Track test status, metrics, and conversation details in real-time.

Use the API to integrate ElevenLabs testing into your workflow.

Prerequisites

Configure ElevenLabs credentials in your agent settings:

Required fields:

Assistant ID: Your ElevenLabs agent ID
ElevenLabs API Key: Your ElevenLabs API key with access to Agents

To find your Assistant ID, go to your ElevenLabs dashboard and navigate to the Agents section. Go to your particular agent you want to test, open that agent and assistant id (11labs agent id) will be visible just below the header.

API Endpoint

POST https://api.cekura.ai/test_framework/v1/scenarios-external/run_scenarios_elevenlabs/

Authentication

Include your API key in the request header:

X-CEKURA-API-KEY: <YOUR_API_KEY>

Request Parameters

scenarios (array | integer | string, required): Test scenarios to run

Array format: List of scenario IDs [123, 456, 789]
Integer format: Run first N scenarios for the agent 5
String format: Run all scenarios for the agent "all"

agent_id (integer, optional): Your Agent ID in Cekura. Required when using integer or “all” format for scenariosfrequency (integer, optional): Number of times to run each scenario (default: 1)name (string, optional): Custom name for this test runconversation_config_override (object, optional): Override conversation configuration for this run

Examples

Single Run (cURL)

curl -X POST \
  'https://api.cekura.ai/test_framework/v1/scenarios-external/run_scenarios_elevenlabs/' \
  -H 'X-CEKURA-API-KEY: <YOUR_API_KEY>' \
  -H 'Content-Type: application/json' \
  -d '{
    "scenarios": [30]
  }'

Multiple Runs (cURL)

curl -X POST \
  'https://api.cekura.ai/test_framework/v1/scenarios-external/run_scenarios_elevenlabs/' \
  -H 'X-CEKURA-API-KEY: <YOUR_API_KEY>' \
  -H 'Content-Type: application/json' \
  -d '{
    "scenarios": [30, 31, 32],
    "frequency": 2,
    "name": "ElevenLabs Test Run"
  }'

Run First N Scenarios (cURL)

curl -X POST \
  'https://api.cekura.ai/test_framework/v1/scenarios-external/run_scenarios_elevenlabs/' \
  -H 'X-CEKURA-API-KEY: <YOUR_API_KEY>' \
  -H 'Content-Type: application/json' \
  -d '{
    "scenarios": 5,
    "agent_id": 123,
    "frequency": 1,
    "name": "First 5 Scenarios"
  }'

Run All Scenarios (cURL)

curl -X POST \
  'https://api.cekura.ai/test_framework/v1/scenarios-external/run_scenarios_elevenlabs/' \
  -H 'X-CEKURA-API-KEY: <YOUR_API_KEY>' \
  -H 'Content-Type: application/json' \
  -d '{
    "scenarios": "all",
    "agent_id": 123,
    "frequency": 1,
    "name": "All Scenarios Test"
  }'

With Custom Configuration (cURL)

curl -X POST \
  'https://api.cekura.ai/test_framework/v1/scenarios-external/run_scenarios_elevenlabs/' \
  -H 'X-CEKURA-API-KEY: <YOUR_API_KEY>' \
  -H 'Content-Type: application/json' \
  -d '{
    "scenarios": [30],
    "conversation_config_override": {
      "tts": {
        "voice_id": "custom_voice_id"
      }
    },
    "frequency": 1,
    "name": "Custom Config Test"
  }'

Python Example

import requests

API_KEY = "<YOUR_API_KEY>"
BASE_URL = "https://api.cekura.ai/test_framework"

headers = {
    "X-CEKURA-API-KEY": API_KEY,
    "Content-Type": "application/json",
}

payload = {
    "scenarios": [30, 31, 32],
    "frequency": 2,
    "name": "ElevenLabs Test Run"
}

resp = requests.post(
    f"{BASE_URL}/v1/scenarios-external/run_scenarios_elevenlabs/",
    headers=headers,
    json=payload,
)
result = resp.json()
print(result)

Python with Configuration Override

import requests

API_KEY = "<YOUR_API_KEY>"
BASE_URL = "https://api.cekura.ai/test_framework"

headers = {
    "X-CEKURA-API-KEY": API_KEY,
    "Content-Type": "application/json",
}

payload = {
    "scenarios": [30],
    "conversation_config_override": {
        "tts": {
            "voice_id": "custom_voice_id"
        }
    },
    "frequency": 1,
    "name": "Custom Config Test"
}

resp = requests.post(
    f"{BASE_URL}/v1/scenarios-external/run_scenarios_elevenlabs/",
    headers=headers,
    json=payload,
)
result = resp.json()
print(result)

Python - Run All Scenarios

import requests

API_KEY = "<YOUR_API_KEY>"
BASE_URL = "https://api.cekura.ai/test_framework"

headers = {
    "X-CEKURA-API-KEY": API_KEY,
    "Content-Type": "application/json",
}

payload = {
    "scenarios": "all",
    "agent_id": 123,
    "frequency": 1,
    "name": "All Scenarios Test"
}

resp = requests.post(
    f"{BASE_URL}/v1/scenarios-external/run_scenarios_elevenlabs/",
    headers=headers,
    json=payload,
)
result = resp.json()
print(result)

Response

{
  "id": 16870,
  "name": "ElevenLabs Test Run",
  "agent": 5,
  "status": "in_progress",
  "success_rate": 0.0,
  "run_as_text": false,
  "is_cronjob": false,
  "runs": [
    {
      "id": 34625,
      "status": "running",
      "scenario": 11547,
      "scenario_name": "Customer Support Scenario",
      "test_profile_data": {"key": "value"}
    }
  ],
  "created_at": "2025-10-16T09:32:59.484534Z",
  "updated_at": "2025-10-16T09:32:59.484942Z"
}

Error Responses

Missing ElevenLabs Credentials

{
  "elevenlabs_creds": ["Please add ElevenLabs credentials"]
}

Missing ElevenLabs Agent ID

{
  "assistant_id": ["ElevenLabs agent ID (assistant_id) is required but not set on the agent."]
}

Invalid Scenarios

{
  "scenarios": ["Invalid scenario IDs or scenarios not found"]
}

Insufficient Balance

{
  "detail": "Insufficient balance for this operation"
}

Monitor Results

Poll for results using the List Runs with IDs API.

Mock Tools (optional)

If your ElevenLabs agent calls tools, you can test it without hitting live services using Mock Tools. Use Auto-Fetch to pull your ElevenLabs tools and generate mock data, then enable the Mock toggle — Cekura updates the webhook tool URLs directly in your agent configuration and restores the originals when you disable the toggle. ElevenLabs’ conversation data already includes the tool calls your agent made, and Cekura extracts those into the transcript automatically. So the Mock Tool Call Accuracy metric works out of the box — unlike a custom integration, you don’t assemble or send the transcript yourself.

Self-improvement

Once your ElevenLabs agent is connected, Improve Agent can fix failing scenarios for you. It works on a private copy of your ElevenLabs agent, edits the copy, re-runs your evaluators, and shows you a before/after per metric. Your live agent is never touched until you click Promote.

Connect the agent

The loop reads and edits the agent through the ElevenLabs API key and Agent ID you configured above, so the Websocket or Chat connection must already be set up. Phone-only connections are not enough — Cekura needs API access to duplicate and edit the agent.

Start from a failure

Launch Improve Agent from a finished run, from a failing category on the Insights page, or by selecting evaluators in the Evaluators tab and using the bulk action.

Review and promote

When the loop finishes, review the per-metric before/after and the exact changes it made. Promote applies them to your live ElevenLabs agent. Discard copy deletes the private copy.

What Cekura can change on an ElevenLabs agent

The system prompt, LLM settings, and tool configuration
The first message and language
Turn-taking settings (turn eagerness, initial wait time)
Speech recognition (ASR) settings
For workflow agents: each subagent node’s prompt and message, and the conditions on edges between nodes

What it never changes

Voice and TTS settings
API keys, webhook URLs, and phone numbers
Transfer destinations, transfer_to_agent tools, and agent-transfer or transfer-to-number workflow nodes

The private copy is created with ElevenLabs’ duplicate API and named with a [selfimprove-clone] suffix, so it is easy to spot in your ElevenLabs dashboard. If your agent transfers callers to other ElevenLabs agents, those agents are cloned too and every transfer is pointed at the corresponding clone, so test calls never reach your live agents. Agents that transfer to more than 10 other agents, or nest transfers more than 3 levels deep, are refused rather than partially cloned. Copies are deleted when you discard them, shortly after you promote, and automatically after about a week if you do neither.

See the full Optimize Agent guide for how the loop reproduces failures, what each outcome means, and troubleshooting.

Get Started

Key Concepts

Guides

Integrations

Advanced

Overview

Configure automatically

Prerequisites

API Endpoint

Authentication

Request Parameters

Examples

Single Run (cURL)

Multiple Runs (cURL)

Run First N Scenarios (cURL)

Run All Scenarios (cURL)

With Custom Configuration (cURL)

Python Example

Python with Configuration Override

Python - Run All Scenarios

Response

Error Responses

Missing ElevenLabs Credentials

Missing ElevenLabs Agent ID

Invalid Scenarios

Insufficient Balance

Monitor Results

Mock Tools (optional)

Self-improvement

Next Steps

​Overview

​Configure automatically

​Prerequisites

​API Endpoint

​Authentication

​Request Parameters

​Examples

​Single Run (cURL)

​Multiple Runs (cURL)

​Run First N Scenarios (cURL)

​Run All Scenarios (cURL)

​With Custom Configuration (cURL)

​Python Example

​Python with Configuration Override

​Python - Run All Scenarios

​Response

​Error Responses

​Missing ElevenLabs Credentials

​Missing ElevenLabs Agent ID

​Invalid Scenarios

​Insufficient Balance

​Monitor Results

​Mock Tools (optional)

​Self-improvement

​Next Steps

Overview

Configure automatically

Prerequisites

API Endpoint

Authentication

Request Parameters

Examples

Single Run (cURL)

Multiple Runs (cURL)

Run First N Scenarios (cURL)

Run All Scenarios (cURL)

With Custom Configuration (cURL)

Python Example

Python with Configuration Override

Python - Run All Scenarios

Response

Error Responses

Missing ElevenLabs Credentials

Missing ElevenLabs Agent ID

Invalid Scenarios

Insufficient Balance

Monitor Results

Mock Tools (optional)

Self-improvement

Next Steps