Skip to content

QUICKSTART.md

Latest commit

 

History

History
122 lines (86 loc) · 3.22 KB

QUICKSTART.md

File metadata and controls

122 lines (86 loc) · 3.22 KB

Quick Start Guide

Get LLM Council Plus running in under 5 minutes.

1. Prerequisites

  • Python 3.10+
  • Node.js 18+
  • uv - Install with: curl -LsSf https://astral.sh/uv/install.sh | sh

2. Install & Run

# Clone the repo
git clone https://github.com/jacob-bd/llm-council-plus.git
cd llm-council-plus

# Install dependencies
uv sync
cd frontend && npm install && cd ..

# Start the app
./start.sh

Open http://localhost:5173 in your browser.

3. First-Time Setup

The Settings panel opens automatically on first launch.

Option A: Use OpenRouter (Easiest)

  1. Get a free API key at openrouter.ai/keys
  2. Paste it in LLM API Keys → OpenRouter
  3. Click Test (auto-saves on success)
  4. Go to Council Config → Select models for your council
  5. Click Save Changes

Option B: Use Ollama (Free & Local)

  1. Install Ollama
  2. Pull a model: ollama pull llama3.1
  3. Start Ollama: ollama serve
  4. In Settings → LLM API Keys → Click Connect for Ollama
  5. Go to Council Config → Enable "Local (Ollama)" → Select models
  6. Click Save Changes

Option C: Use Direct APIs

  1. Get API keys from your preferred providers (OpenAI, Anthropic, Google, etc.)
  2. Enter keys in LLM API Keys → Direct LLM Connections
  3. Click Test for each (auto-saves on success)
  4. Go to Council Config → Enable "Direct Connections" → Select models
  5. Click Save Changes

4. Your First Query

  1. Click + in the sidebar to start a new conversation
  2. Type your question
  3. (Optional) Toggle Web Search for real-time info
  4. Press Enter

Watch as:

  • Stage 1: Each council member responds independently
  • Stage 2: Models anonymously rank each other's responses
  • Stage 3: Chairman synthesizes the final answer

5. Execution Modes

Choose your deliberation depth (toggle in chat header):

Mode What Happens
Chat Only Just Stage 1 - quick individual responses
Chat + Ranking Stages 1 & 2 - see peer rankings
Full Deliberation All 3 stages - complete synthesis (default)

6. Quick Tips

  • Mix model families for diverse perspectives (e.g., GPT + Claude + Gemini)
  • Use Groq for speed - ultra-fast inference
  • Use Ollama for unlimited local queries
  • "I'm Feeling Lucky" button randomizes your council
  • Abort anytime with the stop button in sidebar

Troubleshooting

Problem Solution
Models not appearing Check provider is enabled in Council Config
Rate limit errors Use Groq (14k/day) or Ollama (unlimited)
Port conflict Backend uses 8001, frontend uses 5173
node_modules errors rm -rf frontend/node_modules && cd frontend && npm install

Next Steps

  • Explore System Prompts to customize model behavior
  • Configure Web Search providers (Tavily, Brave) for better results
  • Adjust Temperature sliders for creativity control
  • Export your council config to share or backup

For full documentation, see README.md.

Ask the council. Get better answers.