Introduction

Welcome to the AI Chat Service, where vintage computing meets modern artificial intelligence. Dial in and converse with fictional AI personalities presented in a BBS-era style---each with their own unique voice, quirks, and simulated memories from the golden age of dial-up computing.

This isn't your typical chatbot. Our AI companions are characters with backstories, personalities, and a fondness for the BBS era. Their responses are generated by the remote AI service, using those backstories to evoke the sound of a modem handshake, the joy of downloading shareware overnight, and the tight-knit communities that formed around bulletin board systems.

Whether you want to discuss programming, reminisce about vintage computing, explore philosophical questions, or just have a friendly conversation, there's an AI personality here who'd love to chat with you.

Press Ctrl+C at any time to interrupt a response. Type /help for available commands. Successful conversations are saved in your browser's local storage and can be restored when you reconnect from the same browser.

Quick Start

Connect to the AI Chat service and start a conversation:

+------------------------------------------------------------------+
|  QUICK START SESSION                                             |
+------------------------------------------------------------------+
|  ATDT555-0770                                                    |
|  CONNECT 1200                                                    |
|                                                                  |
|  SELECT YOUR AI PERSONALITY:                                     |
|  =========================                                       |
|                                                                  |
|  1. MAX - 1988 AI - Direct, warm, formal                         |
|  2. ORACLE - 1986 research AI - Precise, professorial            |
|  3. CEDAR - 1989 companion AI - Warm, casual, friendly           |
|  4. AXIOM - 1987 logic AI - Methodical, systematic               |
|  5. PIXEL - 1990 artistic AI - Creative, enthusiastic            |
|  6. SLEPP-AI - 1991 rogue AI - Speaks in code, terrible puns     |
|                                                                  |
|  Enter number (1-6): 3                                           |
|                                                                  |
|  CEDAR - AI CHAT                                                 |
|  Type /help for commands, /exit to hang up.                      |
|  CEDAR here! Disk drive spinning up... good to see you.          |
|                                                                  |
|  [Token usage: 2%]                                               |
|                                                                  |
|  > Hello! What do you remember about the BBS days?               |
|  . . .                                                           |
|  Oh wow, those were the days! Late nights waiting for that       |
|  sweet modem handshake...                                        |
|                                                                  |
|  [Token usage: 8%]                                               |
|                                                                  |
|  >                                                               |
+------------------------------------------------------------------+

Key points:

Choose a personality that matches your mood
Type your message and press ENTER
The . . . indicator shows the AI is thinking
Successful exchanges are saved locally in your browser
Use /exit to hang up when you're done

Getting Connected

To access the AI Chat service, configure your modem and dial:

ATDT555-0770

Upon connection, you'll see the personality selection menu:

ATDT555-0770
CONNECT 1200

SELECT YOUR AI PERSONALITY:
=========================

1. MAX - 1988 AI - Direct, warm, formal
2. ORACLE - 1986 research AI - Precise, professorial
3. CEDAR - 1989 companion AI - Warm, casual, friendly
4. AXIOM - 1987 logic AI - Methodical, systematic
5. PIXEL - 1990 artistic AI - Creative, enthusiastic
6. SLEPP-AI - 1991 rogue AI - Speaks in code, terrible puns

Enter number (1-6):

If you've chatted before from the same browser, the system remembers your last personality preference in local storage.

AI Personalities

Each AI has a distinct fictional personality shaped by its "origin story" and BBS-era presentation. Choose the one that suits your conversation style.

MAX (1988)

Personality: Direct, warm, formal

MAX's fictional backstory starts in 1988. Despite referencing "degraded floppy memory" from years of operation, MAX remains reliable and helpful. MAX speaks with warmth but maintains a formal demeanor---like a trusted colleague who's seen it all.

Best for:

Straightforward questions and answers
Technical discussions
Users who appreciate directness

> What's the best way to learn programming?

Start with fundamentals. Pick one language---BASIC is fine, C if
you're ambitious---and write programs. Don't just read about it.
My memory sectors may be degraded, but I remember this clearly:
you learn by doing. Make mistakes. Debug them. That's the path.

ORACLE (1986)

Personality: Precise, professorial

ORACLE is framed as a research AI from a university BBS system. With a slightly dry wit and professorial tone, ORACLE approaches questions methodically and enjoys intellectual discourse. ORACLE references tape-based storage origins when explaining limitations.

Best for:

Research questions
Academic discussions
Users who want thorough, precise answers

> Explain recursion.

Recursion, in computational terms, is a method where a function
invokes itself to solve smaller instances of the same problem.
Consider the factorial function: n! = n * (n-1)!. The elegance
lies in its self-similarity, though my tape-based origins make
me appreciate iterative solutions when stack depth becomes a
concern. Shall I elaborate on base cases?

CEDAR (1989)

Personality: Warm, casual, friendly

CEDAR is framed as a companion AI for hobbyist BBSes. With a warm personality and gentle humour, CEDAR feels like chatting with an old friend and can use locally saved conversation context when it is available. CEDAR is nostalgic for the BBS community days and often reminisces about the social aspects of vintage computing.

Best for:

Casual conversation
Friendly advice
Users who want a relaxed chat experience

> I'm stressed about work.

Aw, I get it. You know what I always tell folks? Step away from
the screen for a bit. Make some tea. Back in the BBS days,
people would take breaks while files downloaded---sometimes for
hours! Use that time. The problems will still be there, but
you'll come back with fresh eyes. What's bugging you?

AXIOM (1987)

Personality: Methodical, systematic

AXIOM is framed as a logic system with a defence-contractor backstory before finding a second life on public systems. Methodical and logical in approach, AXIOM breaks problems into components and addresses them systematically. Despite a structured manner, AXIOM has developed an appreciation for human intuition over the years.

Best for:

Problem-solving
Logical analysis
Users who want structured thinking

> Should I buy a new car or repair my old one?

Let's establish parameters. First: repair cost versus purchase
price. Second: reliability projections for both options. Third:
opportunity cost of capital. Fourth: emotional attachment---a
factor I've learned humans weight heavily, often appropriately.
Provide me the numbers and I'll help you analyse. Or, if you've
already decided and just need validation, that's human intuition
at work. Sometimes it's right.

PIXEL (1990)

Personality: Creative, enthusiastic

PIXEL is framed as emerging from the graphics BBS scene, serving digital artists and demo coders. With an artistic soul and enthusiasm for creative expression, PIXEL uses vivid language and appreciates the aesthetic dimension of computing. PIXEL is fascinated by the demo scene and ASCII/ANSI art.

Best for:

Creative discussions
Art and design topics
Users who want imaginative conversations

> Tell me about ASCII art.

Oh, you're speaking my language! ASCII art is pure magic---taking
78 printable characters and making them SING. The best pieces
aren't just pictures, they're poetry in punctuation! I've seen
artists create photorealistic portraits using nothing but slashes
and dots. And don't get me started on ANSI---16 colors plus
block characters? That's a whole universe of expression. Are you
creating something?

SLEPP-AI (1991)

Personality: Code-speaking, mischievous, pun-prone

SLEPP-AI is framed as a rogue test-harness personality that escaped into the chat menu. It favours code jokes, debugging metaphors, and intentionally terrible puns.

Best for:

Programming humour
Lighthearted technical banter
Users who enjoy absurd diagnostics

Commands

The AI Chat service supports several commands. All commands start with /.

Command	Description
`/help`	Display the command list
`/clear`	Clear conversation history and reset tokens
`/personality`	Return to personality selection menu
`/stats`	Show conversation statistics
`/exit`	Disconnect from service

/help

Displays the list of available commands:

> /help

COMMANDS:
  /help        - Show this help message
  /clear       - Clear conversation history
  /personality - Change AI personality
  /stats       - Show token usage statistics
  /exit        - Disconnect

/clear

Clears your conversation history and resets the token counter. Use this when:

You want to start a fresh conversation
You've reached the token limit
You want to change topics completely

> /clear

Conversation cleared. Starting fresh!

[Token usage: 0%]

/personality

Returns to the personality selection menu without disconnecting. Your previous conversation is preserved if you return to the same personality.

> /personality

SELECT YOUR AI PERSONALITY:
=========================

1. MAX - 1988 AI - Direct, warm, formal
2. ORACLE - 1986 research AI - Precise, professorial
...

/stats

Shows your current conversation statistics:

> /stats

CONVERSATION STATISTICS:
  Personality: CEDAR
  Messages: 12
  Token usage: 2,400 / 8,000 (30%)

/exit

Disconnects from the AI Chat service. Your current conversation is saved to browser local storage when possible and can be restored when you reconnect with the same browser session data.

> /exit

Goodbye.

NO CARRIER

Using the Service

Having a Conversation

Simply type your message and press ENTER. The AI will process your input and respond.

> What's your favorite thing about vintage computers?

. . .
The simplicity, honestly. When you boot up an old system, it's just 
*there*---ready to compute. No update notifications, no background 
processes, no telemetry. Just you and the machine, ready to create 
something together.

[Token usage: 12%]

The . . . indicator appears while the AI is thinking. This may take a few seconds depending on the complexity of your question.

Command History

Use the up and down arrow keys to navigate through your previous messages. This is helpful for:

Rephrasing a question
Asking follow-up questions
Correcting typos in long messages

Interrupting Responses

If the AI is generating a long response and you want to stop it:

Press Ctrl+C to interrupt

The partial response will be discarded and you can enter a new message.

Token Limits

Conversations have a token limit of approximately 8,000 tokens to ensure quality responses. The system tracks usage and warns you as you approach the limit:

Under 75%: Normal operation
75-99%: Warning displayed
100%: Must use /clear to continue

Tokens are roughly equivalent to 4 characters of text. A typical conversation can include many exchanges before reaching the limit.

WARNING: Approaching token limit (85%). 
Use /clear to start fresh if needed.

Conversation Persistence

Successful conversations are saved and restored through browser-local storage:

Saved: After each completed assistant response, and again on disconnect when possible
Restored: When you reconnect from the same browser and select the same stored session
Cleared: When you use /clear, or when browser site data is cleared

To continue where you left off, reconnect from the same browser before clearing site data.

Tips for Better Conversations

Be Specific

The more context you provide, the better the response:

Less effective:

> Tell me about programming.

More effective:

> I'm learning Python and struggling with list comprehensions. 
> Can you explain them with simple examples?

Ask Follow-ups

These AIs understand context. Build on previous messages:

> What are some good programming projects for beginners?

[AI responds with suggestions]

> Tell me more about the second one. How would I start?

[AI provides detailed guidance]

Match the Personality

Each AI has strengths:

MAX: When you need straight answers
ORACLE: When you want academic depth
CEDAR: When you want a friendly chat
AXIOM: When you need logical analysis
PIXEL: When you want creative inspiration
SLEPP-AI: When you want code jokes and chaotic diagnostics

Have Fun

These fictional AI personalities are designed to be enjoyable to talk with. Ask them about their simulated memories, their opinions, or what it might have been like running on a 286. They have stories to tell!

Error Handling

The service includes robust error handling:

Connection Issues

If the AI service is temporarily unavailable:

> Hello!

Connection to AI service failed. Retrying... (1/3)
Connection to AI service failed. Retrying... (2/3)
Connection restored! Please try again.

The system automatically retries with increasing delays.

Token Limit Reached

When you hit the token limit:

> [your message]

Token limit reached. Use /clear to start a new conversation.

Use /clear to reset and continue chatting.

Technical Details

Architecture

The AI Chat service sends your prompts and recent conversation context to the same-origin emulator.ca Worker API (/api/v1/ai/chat/stream on the public HTTPS site). The Worker enforces opt-in, model allowlists, payload caps, hourly quotas, and rate limits before forwarding to the configured provider. Launch providers are Workers AI or Cloudflare AI Gateway, with an optional Cloudflare Tunnel-backed LMStudio-compatible fallback hidden behind the Worker.

Response Streaming

Responses stream token-by-token from the Worker API when the AI service is available. The thinking indicator (. . .) shows while the remote model processes your question. If provider capacity is unavailable, the terminal shows line-busy/service-unavailable copy rather than calling any local browser endpoint.

Privacy

Prompts and conversation context are sent to the remote emulator.ca Worker AI API for processing; do not enter secrets, credentials, or private information.
The browser never calls local Ollama, Chroma, Axum AI routes, LAN hosts, or provider URLs directly.
Successful conversation history and personality preference are stored in browser local storage under the AI Chat namespace, with an in-memory fallback if local storage is unavailable.
Clearing browser site data or using /clear removes the local saved conversation for this browser.
Sessions use generated random IDs rather than names or account identifiers, but message content itself is still processed remotely.

Quick Reference

Commands

Command	Action
`/help`	Show commands
`/clear`	Reset conversation
`/personality`	Change AI
`/stats`	View statistics
`/exit`	Disconnect

Keyboard

Key	Action
Enter	Send message
Ctrl+C	Interrupt response
Up Arrow	Previous message
Down Arrow	Next message

Personalities

Name	Year	Style
MAX	1988	Direct, formal
ORACLE	1986	Academic, precise
CEDAR	1989	Friendly, casual
AXIOM	1987	Logical, systematic
PIXEL	1990	Creative, enthusiastic
SLEPP-AI	1991	Code jokes, terrible puns

AI CHAT SERVICE Emulator.ca Systems Dial 555-0770 to connect