Skip to content

ai-assist

AI Assistant for Knowledge Workers

An intelligent AI assistant powered by an LLM through the Anthropic Messages API, Skills and MCP (Model Context Protocol) that helps knowledge workers with periodic automated monitoring and interactive querying.

Get Started View on GitHub

Why ai-assist?

  • 🤖 AI-Powered

    Uses an LLM via the Anthropic Messages API for intelligent analysis, with abstract model tiers (low / medium / high) selectable per task.

  • 🔌 MCP Integration

    Connect to any MCP server — DCI, Jira, GitHub, Google Docs, Second Brain — for tools and live data.

  • 💬 Interactive Mode

    A rich terminal UI with streaming responses, markdown rendering, history search, tab completion, and session model/effort controls.

  • 📊 Monitoring

    Automated scheduled checks with smart desktop, file, and console notifications, plus hot reload.

  • 🧠 Knowledge Graph

    A bi-temporal database with vector embeddings tracking entities, relationships, and conversation memory.

  • 🔄 AWL Workflows

    Script multi-step agent workflows with conditionals, loops, variable propagation, and per-task model/effort controls in the Agent Workflow Language.

  • 🚀 Agent Skills

    Install specialized skills following the agentskills.io specification and run them as /<skill> args.

  • âš¡ Event-Driven Actions

    React to the world with triggers from MQTT, D-Bus, and filesystem changes.

Quick start

# Install uv, then clone and sync
git clone https://github.com/ai-assist-org/ai-assist
cd ai-assist
uv sync

# Configure credentials
cp .env.example .env   # edit with your LLM/MCP credentials

# Run
uv run ai-assist                      # interactive mode (default)
uv run ai-assist /monitor             # monitoring mode
uv run ai-assist /query "..."         # one-off query
uv run ai-assist /run workflow.awl    # execute an AWL workflow

See the Getting Started guide for authentication options and next steps.