---
title: "Getting started"
url: "https://raconte.ai/en/docs/getting-started"
---

### Documentation

[Introduction](/en/docs)[Getting started](/en/docs/getting-started)[MCP server](/en/docs/mcp)[Webhooks](/en/docs/webhooks)[REST API](/en/docs/api)[CLI](/en/docs/cli)[SDK](/en/docs/sdk)[Agent skill](/en/docs/skill)

### Guides

[Configure the MCP server](/en/guides/mcp-setup)[Using the MCP server](/en/guides/mcp-usage)

[Raconte](/) ⟩ [Documentation](/en/docs)

# Getting started

Create an account, send your first interview, read the transcript.

## Create your account

1.  Go to [raconte.ai](https://raconte.ai) and sign up.
2.  Confirm your email address.
3.  You land in an organization. Every interview, invitation and API key is scoped to that organization.

## Send your first interview

1.  Open **Interviews** and click **New interview**.
2.  Write the **prompt**. Tell the AI what to ask: subject, tone, the points it must cover. A few sentences are enough.
3.  Pick the **language**. The AI will speak and the messages will be transcribed in that locale.
4.  Optionally tweak the **title**, **intro** (what the invitee reads before starting), and **first message** (the AI’s opening line). The AI generates sensible defaults if you skip them.
5.  Add a recipient by email or phone, or just save and copy the share link to send it manually.

Each send produces an **invitation** with its own slug-URL. You can send the same interview to multiple people: each one gets their own invitation, their own transcript, their own sentiment analysis.

## Read the transcript

Open an interview to see every invitation it has produced. A green checkmark means the call completed and the transcript is ready.

A transcript contains:

*   The full message log (assistant and user turns).
*   A per-message sentiment (positive, neutral, negative) on the user turns.
*   A short summary of what the respondent actually said. Greetings, fillers, and meta requests are filtered out.
*   The audio recording, message by message, when the invitee gave microphone permission.

## What to set up next

*   **Voice and system prompt**. From **Settings → Interviews**, set an organization-wide system prompt (style, persona, things to always say or avoid). It is prepended to every interview unless you disable it on a specific one.
*   **Webhooks**. Get a POST request when an interview starts and when it ends. See [Webhooks](/en/docs/webhooks).
*   **MCP server**. Create interviews, manage invitations and read transcripts from Claude, Cursor or Opencode. See [MCP server](/en/docs/mcp) for the API, [Set up the MCP server](/en/guides/mcp-setup) for the per-client configuration.
*   **CLI and SDK**. Drive interviews and invitations from the terminal with [`@raconte/cli`](/en/docs/cli), or from a TypeScript or Node app with [`@raconte/node-sdk`](/en/docs/sdk).
*   **Billing**. Top up the wallet from **Settings → Billing**. Calls are billed by the minute of active speech.

**ℹ️ Free monthly minutes**

Every organization gets a few free minutes per month. You can run a complete interview without topping up first; the wallet only kicks in once you exceed the free allowance.

Table of contents

[1\. Create your account](#create-your-account)[2\. Send your first interview](#send-your-first-interview)[3\. Read the transcript](#read-the-transcript)[4\. What to set up next](#what-to-set-up-next)
