Live Demos - Mnemosyne

Demo 1: AI Debate Moderator

Use Case: Tracking multi-participant conversations. Mnemosyne uses the speaker field to remember who said what, allowing the AI to act as an intelligent moderator.

> Moderator, Alice, and Bob join the debate room (userId: 1)...

POST /v1/chat/completions
{
  "userId": 1,
  "messages": [
    { "speaker": "Alice", "content": "AI will create new job categories." },
    { "speaker": "Bob", "content": "The scale of automation is unprecedented." }
  ]
}

> Moderator asks for a summary...

{ "role": "assistant", "content": "Alice argues for job creation, while Bob is concerned about job displacement." }

Demo 2: Persistent Support Bot

Use Case: Remembering user history across sessions. Mnemosyne links all interactions to a persistent userId, providing seamless customer support.

> Day 1: User reports an issue...

POST /v1/chat/completions
{
  "userId": 2,
  "messages": [{ "role": "user", "content": "My X-1000 router is not connecting." }]
}

AI Response: "I've logged the issue for your X-1000 router. We are looking into it." (Memory stored)

> Day 3: User follows up...

POST /v1/chat/completions
{
  "userId": 2,
  "messages": [{ "role": "user", "content": "Any update on my issue?" }]
}

{ "role": "assistant", "content": "Yes, I see you're having an issue with your X-1000 router. Our team has pushed a firmware update. Have you tried rebooting it?" }

Demo 3: Card Game Analyst

Use Case: Storing and reasoning over structured, non-dialogue data using the updateMemory tool.

> Round 1 (Flop): Store public cards...

POST /v1/memory
{
  "userId": 3, // The Game ID
  "memories": [{
    "key": "round_1_flop",
    "data": { "cards": ["Ace of Spades", "10 of Clubs", "2 of Hearts"] }
  }]
}

> Ask AI for analysis...

POST /v1/chat/completions
{
  "userId": 3,
  "messages": [{ "role": "user", "content": "What are the possible straights?" }]
}

{ "role": "assistant", "content": "Based on the flop (A, 10, 2), possible straight draws include King-Queen-Jack, Jack-Queen-King, and 3-4-5." }

Demo 4: Dynamic Workflow Agent

Use Case: Storing and iteratively updating complex JSON objects, perfect for AI agent state management.

> User asks to plan a trip...

POST /v1/chat/completions
{ "userId": 4, "messages": [{ "role": "user", "content": "Plan a 2-day trip to Paris." }] }

AI generates a plan and saves it to memory with key "paris_trip_plan".

> User requests a change...

POST /v1/chat/completions
{
  "userId": 4,
  "messages": [{ "role": "user", "content": "Actually, let's add a visit to the Louvre on Day 1." }]
}

AI retrieves the "paris_trip_plan" memory, updates the JSON object with the new activity, and saves it back.

{ "role": "assistant", "content": "Great! I've updated your itinerary to include the Louvre on Day 1." }

Demo 5: Data Analysis Assistant

Use Case: Acting as a memory backend for Retrieval-Augmented Generation (RAG) systems.

> Store quarterly sales data...

POST /v1/memory
{
  "userId": 5, // The Analysis Session ID
  "memories": [
    { "key": "sales_q1", "data": { "quarter": "Q1", "revenue": 110000 } },
    { "key": "sales_q2", "data": { "quarter": "Q2", "revenue": 150000 } }
  ]
}

> Ask a question in natural language...

POST /v1/chat/completions
{
  "userId": 5,
  "messages": [{ "role": "user", "content": "Compare Q1 and Q2 sales." }]
}

{ "role": "assistant", "content": "Q2 sales were $150,000, which is an increase of $40,000 over Q1's $110,000." }