Chabeau Server

A lightweight Node.js server providing an ELIZA chatbot through an OpenAI-compatible API. Perfect for testing chatbot applications without requiring an actual LLM.

Quick Start

npm install
npm start

Server runs on http://localhost:3000 (set PORT environment variable to customize).

API Endpoints

Chat Completions

POST /v1/chat/completions - OpenAI-compatible chat endpoint

Authentication: Authorization: Bearer <any-token> (use "badtoken" to test auth failures)

Request:

{
  "model": "eliza-1.0",
  "messages": [{"role": "user", "content": "Hello!"}],
  "stream": false
}

Response:

{
  "id": "chatcmpl-1234567890",
  "object": "chat.completion",
  "created": 1677610602,
  "model": "eliza-1.0",
  "choices": [{
    "index": 0,
    "message": {
      "role": "assistant",
      "content": "Hello... I'm glad you could drop by today."
    },
    "finish_reason": "stop"
  }],
  "usage": {
    "prompt_tokens": 10,
    "completion_tokens": 8,
    "total_tokens": 18
  }
}

Other Endpoints

GET /v1/models - List available models
GET /health - Health check

Testing Features

Long Response Generation

! - Single long paragraph (~85 tokens)
!! - Two paragraphs (~170 tokens)
!!! - Three paragraphs (~260 tokens)

Perfect for testing "re-generate reply" functionality with varied content.

Code Block Generation

@ - Single random code block in various programming languages
@@ - Three random code blocks in various programming languages
@@@ - Five random code blocks in various programming languages

Each code block is formatted as a fenced markdown code block with a brief explanation text. Supported languages include Python, SQL, Bash, JavaScript, HTML, CSS, Java, C++, Go, and Rust.

Streaming Speed Control

> / >> / >>> - Increase speed (5x/10x/100x baseline)
< / << / <<< - Decrease speed / reset to baseline (50 tokens/sec)

Request Latency Control

) / )) / ))) - Add latency (500ms/2.5s/5s)
( / (( / ((( - Reduce latency / reset to baseline (50ms)

Markdown Table Generation

| - Single random markdown table with varied content
|| - 2-3 random markdown tables with varied content
||| - 3-4 random markdown tables with varied content

Each table includes different markdown formatting features like complex content, code snippets, links, line breaks, and various table structures for comprehensive testing.

Link Generation

% - Simple text with plain URLs and basic markdown links
%% - Multiple paragraphs with various link formats and styles
%%% - Complex content including markdown tables with embedded links

Generates content with links in various formats (plain URLs, markdown links in paragraphs, and links within tables) with flavor text inspired by the intersection of Douglas Adams' wit, Ted Nelson's Xanadu vision, and Tim Berners-Lee's web invention.

Rich Markdown Paragraphs

$ - A single wry markdown paragraph about money, adorned with visuals
$$ - Two rich paragraphs with imagery, horizontal rules, and emojis
$$$ - Three richly formatted paragraphs for maximum markdown variety

These responses focus on reflective, tongue-in-cheek commentary about money—no tables or code fences—while showcasing markdown images, horizontal rules, and expressive emoji.

Terminal-Breaking Emoji

& - Single file with complex emoji modifiers (skin tones, ZWJ sequences)
&& - Two files with diverse emoji combinations
&&& - Three files with maximum emoji complexity

Content featuring human-animal relationships with extensive use of emoji modifiers including skin tone variations (👨🏾‍💼 👩🏻‍🔬), gender+profession ZWJ sequences (🐕‍🦺 👨🏽‍⚕️), multi-person families (👨‍👩‍👧‍👦), keycap sequences (#️⃣ *️⃣), flag tags (🏴󠁧󠁢󠁳󠁣󠁴󠁿), and variation selectors (☀️). Includes tables, formatted prose, and unformatted prose to stress-test terminal rendering.

International Text

~ - Single file with CJK, Arabic, or mixed scripts
~~ - Two files with diverse language combinations
~~~ - Three files with maximum script complexity

Multilingual content testing complex text rendering including CJK (Chinese 中文, Japanese 日本語, Korean 한국어), RTL scripts (Arabic العربية, Hebrew עברית), South Asian scripts (Devanagari हिन्दी, Bengali বাংলা), Southeast Asian (Thai ไทย, Khmer ភាសាខ្មែរ), Cyrillic (Русский), and others. Features bidirectional text, multiple number systems (١٢٣ ๑๒๓ १२३), currency symbols (﷼ ₪ ₹ ฿), and complex tables mixing scripts.

API Debugging Commands

system - Returns the system prompt from the API request (if any) in a markdown code block
raw - Returns the complete raw API payload as formatted JSON in a markdown code block

These commands help debug API integrations by showing exactly what data was sent to the server. If no system prompt was provided, the system command returns "No system prompt was submitted."

Example Usage

cURL

# Basic chat
curl -X POST http://localhost:3000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer test-token" \
  -d '{"model": "eliza-1.0", "messages": [{"role": "user", "content": "Hello!"}]}'

# Streaming
curl -X POST http://localhost:3000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer test-token" \
  -d '{"model": "eliza-1.0", "messages": [{"role": "user", "content": "Hello!"}], "stream": true}'

JavaScript

const response = await fetch('http://localhost:3000/v1/chat/completions', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'Authorization': 'Bearer test-token'
  },
  body: JSON.stringify({
    model: 'eliza-1.0',
    messages: [{ role: 'user', content: 'Hello!' }]
  })
});

const data = await response.json();
console.log(data.choices[0].message.content);

ELIZA Patterns

The chatbot recognizes classic ELIZA patterns:

Needs: "I need..." → "Why do you need...?"
Feelings: "I feel..." → "Tell me more about these feelings."
Questions: "Why..." → "Why don't you tell me the reason why...?"
Statements: "I am..." → "How long have you been...?"

Server Development

Content Management System

The server uses a filesystem-based content management system that allows server developers to easily add new test content without modifying the core bot logic.

Directory Structure

test-content/
├── config.toml                 # Main configuration file
├── code-blocks/                # Programming language examples
│   ├── python/                 # Python code samples
│   ├── javascript/             # JavaScript code samples
│   ├── sql/                    # SQL query examples
│   ├── bash/                   # Shell script examples
│   ├── html/                   # HTML markup examples
│   ├── css/                    # CSS styling examples
│   ├── java/                   # Java code examples
│   ├── cpp/                    # C++ code examples
│   ├── go/                     # Go code examples
│   └── rust/                   # Rust code examples
├── long-responses/             # Extended paragraph content
├── rich-markdown/              # Rich markdown paragraphs with imagery
├── tables/                     # Markdown table examples
├── links/                      # Content with embedded links
├── terminal-breaking-emoji/    # Complex emoji with modifiers
└── international-text/         # Multilingual and multi-script content

Adding New Content

Code Blocks: Add new files to the appropriate language directory under test-content/code-blocks/. Files should contain complete, runnable code examples with proper file extensions (.py, .js, .sql, etc.).

Long Responses: Add .txt files to test-content/long-responses/. Each file should contain one coherent paragraph of 80-120 words.

Rich Markdown: Add .md files to test-content/rich-markdown/. Each file should feature expressive paragraphs without tables or code fences, optionally including horizontal rules, images, and emoji for visual richness.

Tables: Add .md files to test-content/tables/. Each file should contain one complete markdown table with headers and data.

Links: Add content files to test-content/links/. Support plain URLs, markdown links, and table-embedded links.

Terminal-Breaking Emoji: Add .md files to test-content/terminal-breaking-emoji/. Content should include complex emoji with skin tone modifiers, ZWJ sequences, multi-person groupings, keycap sequences, and flag tags to stress-test terminal rendering.

International Text: Add .md files to test-content/international-text/. Include content in various scripts (CJK, Arabic, Hebrew, Devanagari, Thai, Cyrillic, etc.) with bidirectional text, multiple number systems, and mixed-script tables.

Configuration

The test-content/config.toml file controls content categories, command mappings, and system settings. The system automatically discovers new content files without requiring configuration changes.

File Organization Standards

Use descriptive filenames that indicate the content purpose
Maintain consistent file extensions for each content type
Organize code blocks by programming language
Keep content files focused on single examples or concepts
Ensure all content is properly formatted and readable

Project Structure

The server is organized into focused, single-responsibility modules:

index.js - Main entry point, starts the server
server.js - HTTP server logic and API endpoints
eliza-bot.js - ELIZA chatbot implementation using ContentManager
lib/content-manager.js - Content loading and management system
lib/config-loader.js - TOML configuration file handling
test-content/ - Filesystem-based content organization
- config.toml - Content system configuration
- code-blocks/ - Programming language examples by language
- long-responses/ - Extended paragraph content files
- rich-markdown/ - Rich markdown paragraphs with imagery and emoji
- tables/ - Markdown table examples
- links/ - Content with various link formats
- terminal-breaking-emoji/ - Complex emoji with modifiers and ZWJ sequences
- international-text/ - Multilingual content with diverse scripts

Development

Dependencies

The server uses minimal dependencies:

smol-toml - TOML configuration file parsing
Node.js built-in modules (http, url, fs, path)

Modifying Content

ELIZA Patterns: Edit the patterns array in the ElizaBot class within eliza-bot.js.

Test Content: Add files to appropriate directories under test-content/. The system automatically discovers and loads new content.

API Endpoints: Modify the request handling logic in server.js.

Server Configuration: Update the startup logic in index.js.

Content System: Modify lib/content-manager.js for content loading behavior or test-content/config.toml for content settings.

License

CC0-1.0 - Public Domain

Name		Name	Last commit message	Last commit date
Latest commit History 29 Commits
.github/workflows		.github/workflows
lib		lib
test-content		test-content
tests		tests
.gitignore		.gitignore
AGENTS.md		AGENTS.md
CONTENT_MANAGEMENT.md		CONTENT_MANAGEMENT.md
LICENSE		LICENSE
README.md		README.md
eliza-bot.js		eliza-bot.js
index.js		index.js
package-lock.json		package-lock.json
package.json		package.json
server.js		server.js
vitest.config.js		vitest.config.js

License

permacommons/chabeau-server

Folders and files

Latest commit

History

Repository files navigation

Chabeau Server

Quick Start

API Endpoints

Chat Completions

Other Endpoints

Testing Features

Long Response Generation

Code Block Generation

Streaming Speed Control

Request Latency Control

Markdown Table Generation

Link Generation

Rich Markdown Paragraphs

Terminal-Breaking Emoji

International Text

API Debugging Commands

Example Usage

cURL

JavaScript

ELIZA Patterns

Server Development

Content Management System

Directory Structure

Adding New Content

Configuration

File Organization Standards

Project Structure

Development

Dependencies

Modifying Content

License

About

Topics

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Languages

Packages