Agent playground simulation

Let an AI-simulated user talk to an agent and stream the conversation.

POSThttps://api.tried.com.au/agent-playground-simulation/{agent_id}

Runs a whole test conversation between a voice agent's latest version and a simulated user that follows your persona prompt. Each transcript item is streamed as a server-sent event. Nothing is stored. Beyond Retell

With an LLM key the simulated user is an LLM. In mock mode the persona prompt is turned into a scripted list of first-person lines. The member role may call this route.

Request#

Path parameters#

Parameter	Type	Required	Description
`agent_id`	string	Yes	The voice agent to test.

Body#

Field	Type	Required	Description
`user_prompt`	string	Yes	Persona and goal of the simulated user. 1 to 8000 characters.
`dynamic_variables`	object	No	String, number or boolean values (converted to strings). Default `{}`.
`tool_mocks`	object[]	No	Up to 50 `{ "tool_name", "output" }` returned instead of running the tool.
`llm_model`	string	No	Model for the simulated user. Default `gpt-5.6-terra`.
`max_utterances`	integer	No	2 to 400. Default `40`.
`start_component_id`	string	No	Conversation flow only: run inside one component, Begin node to Exit node.

Example request#

curl -N -X POST "https://api.tried.com.au/agent-playground-simulation/agent_3f9a1c0b2d4e5f6a7b8c9d0e" \
  -H "Authorization: Bearer $TRIED_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"user_prompt":"You are Sam. You want to move your Thursday booking to Friday morning.","max_utterances":20}'

const res = await fetch("https://api.tried.com.au/agent-playground-simulation/agent_3f9a1c0b2d4e5f6a7b8c9d0e", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.TRIED_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({ user_prompt: "You are Sam. You want to move your Thursday booking to Friday morning.", max_utterances: 20 }),
});
console.log(await res.text());

import json, os, requests

res = requests.post(
    "https://api.tried.com.au/agent-playground-simulation/agent_3f9a1c0b2d4e5f6a7b8c9d0e",
    headers={"Authorization": f"Bearer {os.environ['TRIED_API_KEY']}"},
    json={"user_prompt": "You are Sam. You want to move your Thursday booking to Friday morning.", "max_utterances": 20},
    stream=True,
)
for line in res.iter_lines(decode_unicode=True):
    if line.startswith("data: "):
        print(json.loads(line[6:]))

Response#

200 OK with Content-Type: text/event-stream.

data: {"type":"item","item":{"role":"agent","content":"Hi, thanks for calling Bright Dental."}}

data: {"type":"item","item":{"role":"user","content":"Hi, I need to move my Thursday booking."}}

data: {"type":"done","ended_by":"agent_ended","utterances":8}

`type`	Fields
`item`	`item`: a transcript item with `role` `agent`, `user`, `tool_call_invocation` (`tool_call_id`, `name`, `arguments`), `tool_call_result` (`tool_call_id`, `content`) or `node_transition` (`former_node_id`, `new_node_id`).
`done`	`ended_by` (`user_ended`, `agent_ended`, `script_finished`, `utterance_limit`, `time_limit`, `aborted`) and `utterances` (agent and user turns).
`error`	`message`.

Errors#

Status	When
`400`	Validation failed, or `start_component_id` is used with an agent that is not a conversation flow, or the component has no nodes.
`401`	Missing or invalid API key.
`403`	The key's IP allowlist or scopes reject the request. This route needs the `agents` scope.
`404`	The agent or component does not exist.
`429`	Rate limit exceeded. See Rate limits.

See Errors for the body format.