---
title: "WhisperChat Developer Portal | API, MCP Server & OpenAPI Docs"
description: "Build on WhisperChat: REST API reference, OpenAPI specification, MCP server, agent authentication, integration guides and machine-readable docs — every developer resource in one place."
canonical: "https://www.whisperchat.ai/developers"
last-updated: "2026-10-22"
source: "https://www.whisperchat.ai/developers"
---

# WhisperChat Developer Portal | API, MCP Server & OpenAPI Docs

> Build on WhisperChat: REST API reference, OpenAPI specification, MCP server, agent authentication, integration guides and machine-readable docs — every developer resource in one place.

## Build on WhisperChat

WhisperChat is an AI chatbot for website customer support, and everything the product does from its own dashboard can be done from your code: create a chatbot, feed it knowledge, train it, ask it questions, read the conversations back. Two ways in — a REST API and an MCP server — described by the same set of documents below.

**REST base URL:** `https://api.whisperchat.ai` (23 operations under `/v1`)

**MCP endpoint:** `https://www.whisperchat.ai/mcp` (Streamable HTTP)

**Authentication:** API key as a bearer token for REST, OAuth 2.1 for MCP. API access starts on the $49 Standard plan; the free plan needs no card.

## WhisperChat developer resources

Every reference document, at a stable URL. Nothing here needs an account to read.

[

REST API documentation

### WhisperChat API reference

Every endpoint of the WhisperChat REST API (https://api.whisperchat.ai, version 1.0.0): chatbots, knowledge sources, training and chat sessions, with authentication and examples.

](https://www.whisperchat.ai/docs/api)[

Machine-readable spec

### WhisperChat OpenAPI specification

The OpenAPI 3 document the reference is generated from. Feed it to a code generator, an HTTP client, or an LLM function-calling adapter.

](https://www.whisperchat.ai/openapi.json)[

Model Context Protocol

### WhisperChat MCP server

Streamable HTTP at https://www.whisperchat.ai/mcp with 21 tools. OAuth 2.1 with PKCE and Dynamic Client Registration; initialize, tools/list and resources/list answer without a token so you can inspect it first.

](https://www.whisperchat.ai/docs/mcp)[

Model Context Protocol (read-only)

### WhisperChat docs MCP server

POST MCP to /docs/mcp for search\_docs, fetch\_doc and list\_docs over this site's own index. No account, no auth — for agents answering questions about WhisperChat.

](https://www.whisperchat.ai/docs/mcp)[

Auth documentation

### WhisperChat agent authentication

How an agent registers and obtains credentials: auth.md, RFC 9728 protected-resource metadata, RFC 8414 authorization-server metadata. Scopes: chatbot:read, chatbot:write, chatbot:delete, knowledge:write.

](https://www.whisperchat.ai/auth.md)[

Integration documentation

### WhisperChat integration guides

WordPress, Discord, Telegram, Slack, Shopify and Odoo, plus the embed snippet for any website.

](https://www.whisperchat.ai/docs/integration)

## WhisperChat API quickstart

Three calls: authenticate, find a chatbot, ask it something. Full parameters are in the [API reference](https://www.whisperchat.ai/docs/api).

REST

```
# 1. Create an API key in the dashboard: Account -> API token
export WHISPERCHAT_API_KEY=wc_...

# 2. List your chatbots
curl https://api.whisperchat.ai/v1/chatbots \
  -H "Authorization: Bearer $WHISPERCHAT_API_KEY"

# 3. Ask one of them a question
curl https://api.whisperchat.ai/v1/chatbot/<chatbotId>/chat \
  -H "Authorization: Bearer $WHISPERCHAT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"message": "What are your opening hours?"}'
```

Or connect an MCP client — Claude, ChatGPT, Cursor, VS Code — and skip the HTTP entirely:

MCP client config

```
{
  "mcpServers": {
    "whisperchat": {
      "type": "http",
      "url": "https://www.whisperchat.ai/mcp"
    }
  }
}
```

## WhisperChat machine-readable surfaces

For agents reading this site rather than people. All unauthenticated.

-   [llms.txt](https://www.whisperchat.ai/llms.txt)The whole product as one plaintext document.
-   [docs/llms.txt](https://www.whisperchat.ai/docs/llms.txt)Documentation-scoped index.
-   [api/llms.txt](https://www.whisperchat.ai/api/llms.txt)REST + MCP scoped index.
-   [/ask](https://www.whisperchat.ai/ask?query=how+do+i+train+a+chatbot)NLWeb endpoint: ask a question, get JSON or SSE.
-   [Markdown twins](https://www.whisperchat.ai/developers.md)Append .md to any page, or send Accept: text/markdown.
-   [.well-known/mcp.json](https://www.whisperchat.ai/.well-known/mcp.json)MCP server card.
-   [.well-known/ard.json](https://www.whisperchat.ai/.well-known/ard.json)Agentic resource catalog.
-   [.well-known/api-catalog](https://www.whisperchat.ai/.well-known/api-catalog)RFC 9727 linkset.