Skip to content

Ajayduddi/AutoPilot

BranchesTags

Repository files navigation

AutoPilot

Chat-First Automation Platform
Conversational control center for workflows, agents, and automations

Docker License

📋 Overview

AutoPilot transforms how you interact with automation workflows. Instead of manual triggers and complex dashboards, simply chat with your automations. It connects to multiple workflow providers and LLMs, giving you a unified conversational interface for execution, monitoring, and approval workflows.

✨ Features

🤖 AI-Powered Chat Interface

  • Natural Language Control - Trigger and manage workflows through conversation
  • Streaming Responses - Real-time SSE streaming for assistant replies
  • Rich Message Blocks - Render summaries, email drafts, questions, and source references
  • Context Memory - Intelligent conversation context across threads

🔌 Multi-Provider Workflow Support

Connect to your existing automation infrastructure:

  • n8n - Self-hosted or cloud n8n instances
  • Zapier - Zapier webhooks and Zaps
  • Make (Integromat) - Scenario execution
  • Custom HTTP - Generic webhook endpoints
  • Simulator - Test workflows without external dependencies

🧠 Flexible LLM Integration

Switch between providers based on your needs:

  • OpenAI (GPT-4o, GPT-4o-mini)
  • Google Gemini (Gemini 2.5 Flash)
  • Ollama - Self-hosted models (Llama, Mistral, etc.)
  • Groq - Fast inference
  • Mistral - European AI models

🔔 Real-Time Notifications

  • Web Push - Browser notifications for workflow events
  • SSE Streams - Live updates in the chat interface
  • Approval Requests - In-app and push notification approvals

📎 File Attachments

  • Upload documents for AI processing
  • Automatic text extraction and chunking
  • MIME type validation and virus scanning (ClamAV)
  • Structured metadata extraction

🔐 Security-First

  • Google OAuth + session-based authentication
  • Encrypted API key storage (AES-256-GCM)
  • CSRF protection, rate limiting, security headers
  • Webhook signature verification

🚀 Quick Start

Prerequisites

  • Docker installed
  • PostgreSQL 16+ (or use Docker Compose with included services)

Docker Run

# Pull the image
docker pull ajayduddi/autopilot:latest

# Copy and configure environment variables
cp .env.example .env
# Edit .env with your actual values (database, secrets, callbacks)

# Run with environment file
docker run -d \
  --name autopilot \
  -p 3000:3000 \
  --env-file .env \
  ajayduddi/autopilot:latest

Environment Configuration

Create a .env file with the following (copy from .env.example):

PORT=3000
DATABASE_URL=postgres://postgres:postgres@localhost:5432/autopilot
FRONTEND_ORIGIN=https://your-domain.com
AUTOPILOT_HOME=/home/<your-username>/.autopilot
NODE_ENV=production
VITE_API_URL=https://your-backend-domain.com

# ─── Authentication / Sessions ──────────────────────────────
AUTH_COOKIE_SECRET=change-this-secret
SESSION_TTL_DAYS=14
AUTH_REVOKE_OTHER_SESSIONS_ON_PASSWORD_CHANGE=true
GOOGLE_CLIENT_ID=
GOOGLE_CLIENT_SECRET=
GOOGLE_REDIRECT_URI=https://your-domain.com/api/auth/google/callback

# ─── Workflow / Webhook Integration ─────────────────────────
# All workflow execution endpoints are configured per workflow in the UI.
# Optional fallback secret for provider callbacks to /api/webhooks/*.
# Preferred: generate/revoke keys from Settings > Webhook Callbacks.
WEBHOOK_CALLBACK_SECRET=change-me
CALLBACK_BASE_URL=https://your-domain.com

# ─── LLM Provider Defaults ──────────────────────────────────
PROVIDER_API_KEY_ENCRYPTION_KEY=change-this-32-byte-secret
CLAMAV_HOST=127.0.0.1
CLAMAV_PORT=3310
ATTACHMENT_SCAN_HTTP_URL=
ATTACHMENT_SCAN_HTTP_TOKEN=

# ─── Stability Rollout Flags (Compatibility-First) ──────────
# Keep disabled by default to preserve current behavior.
FEATURE_TYPED_CONTRACTS=false
FEATURE_STRUCTURED_LOGGING=false

# ─── Web Push (WhatsApp/Telegram-style browser push) ────────
# Generate once and keep stable in production:
#   npx web-push generate-vapid-keys
VAPID_PUBLIC_KEY=
VAPID_PRIVATE_KEY=
VAPID_SUBJECT=mailto:example@example.com

Docker Compose

Create a docker-compose.yml file:

version: '3.8'

services:
  autopilot:
    container_name: autopilot
    image: ajayduddi/autopilot:latest
    restart: always
    env_file: .env.docker
    environment:
      - DATABASE_URL=postgresql://${POSTGRES_USER:-postgres}:${POSTGRES_PASSWORD:-postgres}@db:5432/${POSTGRES_DB:-autopilot}
    depends_on:
      - db
    command: >
      sh -c "bun /app/apps/backend/dist/db/migrate.js &&
             bun /app/apps/backend/dist/index.js"
    ports:
      - "3000:3000"
    volumes:
      - autopilot_data:/data

  db:
    image: postgres:16-alpine
    environment:
      - POSTGRES_USER=postgres
      - POSTGRES_PASSWORD=postgres
      - POSTGRES_DB=autopilot
    volumes:
      - postgres_data:/var/lib/postgresql/data

volumes:
  autopilot_data:
  postgres_data:

Then run:

# Create compose-specific env file
cp .env.example .env.docker

# Keep DB credentials in one place for both services
echo "POSTGRES_USER=postgres" >> .env.docker
echo "POSTGRES_PASSWORD=postgres" >> .env.docker
echo "POSTGRES_DB=autopilot" >> .env.docker

# Remove DATABASE_URL from .env.docker if present;
# compose builds it from POSTGRES_* and service hostname 'db'.

# Start services
docker compose up -d

Note on Single-Server Architecture:

In Docker, AutoPilot runs as a single process where the backend serves the static frontend directly. Even though both run on the same server, you must still configure:

Variable Why Required
FRONTEND_ORIGIN OAuth callbacks (Google login redirects) and CORS headers
CALLBACK_BASE_URL Webhook callbacks from workflow providers

Example production setup:

# Your public domain
FRONTEND_ORIGIN=https://your-domain.com
CALLBACK_BASE_URL=https://your-domain.com
GOOGLE_REDIRECT_URI=https://your-domain.com/api/auth/google/callback

⚙️ Configuration

Required Environment Variables

Variable Description
DATABASE_URL PostgreSQL connection string
AUTH_COOKIE_SECRET Secret for session cookie signing (32+ chars)
PROVIDER_API_KEY_ENCRYPTION_KEY Key for encrypting stored API keys (32+ chars)
CALLBACK_BASE_URL Public URL for webhook callbacks

Optional Variables

Variable Default Description
PORT 3000 Server port
FRONTEND_ORIGIN - CORS origin for frontend
GOOGLE_CLIENT_ID - Google OAuth client ID
GOOGLE_CLIENT_SECRET - Google OAuth client secret
WEBHOOK_CALLBACK_SECRET - Fallback webhook verification secret
VAPID_PUBLIC_KEY - Web Push public key
VAPID_PRIVATE_KEY - Web Push private key
VAPID_SUBJECT - Web Push contact email

Runtime Config

Advanced configuration via ~/.autopilot/config.json:

{
  "uploadDir": "./uploads",
  "approvalMode": "default",
  "DEFAULT_TIMEZONE": "UTC",
  "ALLOW_PRIVATE_MODEL_FETCH": true,
  "MODEL_FETCH_TIMEOUT_MS": 10000,
  "MAX_MODEL_FETCH_BYTES": 2097152,
  "OLLAMA_URL": "http://localhost:11434",
  "CONTEXT_MODE_ENABLED": true,
  "CONTEXT_MODE_DEBUG": false,
  "CONTEXT_MODE_MAX_RETRIEVAL": 24,
  "CONTEXT_MODE_MODEL_MAX_RETRIEVAL_JSON": {
    "gpt-4o": 24,
    "gpt-4o-mini": 32,
    "gemini-1.5-pro": 40,
    "llama3": 20,
    "mistral": 28,
    "*": 24
  },
  "CONTEXT_MODE_CONTENT_MAX_LEN": 120000,
  "CONTEXT_MODE_SUMMARY_MAX_LEN": 300,
  "CONTEXT_MODE_TARGET_WINDOW_TOKENS": 250000,
  "CONTEXT_MODE_HISTORY_BUDGET_TOKENS": 160000,
  "CONTEXT_MODE_RETRIEVED_CONTEXT_BUDGET_TOKENS": 70000,
  "CONTEXT_MODE_MAX_MESSAGE_TOKENS": 12000,
  "CONTEXT_MODE_MAX_CONTEXT_ITEM_TOKENS": 18000,
  "CONTEXT_MODE_CACHE_DATA_BUDGET_TOKENS": 48000,
  "CONTEXT_MODE_INDEX_WORKFLOW_RUNS": true,
  "CONTEXT_MODE_INDEX_DECISIONS": true,
  "CONTEXT_MODE_INDEX_THREAD_STATE": true,
  "CONTEXT_MODE_TTL_DAYS": 30,
  "CONTEXT_MODE_CACHE_ANSWER": true,
  "CONTEXT_MODE_CACHE_STALE_MINS": 15,
  "MAX_UPLOAD_MB": 25,
  "MAX_FILES_PER_MESSAGE": 6,
  "ALLOWED_MIME_TYPES": [
    "image/*",
    "audio/*",
    "text/*",
    "application/json",
    "application/xml",
    "application/pdf",
    "application/msword",
    "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
    "application/vnd.ms-excel",
    "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
    "application/vnd.ms-powerpoint",
    "application/vnd.openxmlformats-officedocument.presentationml.presentation"
  ],
  "ATTACHMENT_PROCESS_TIMEOUT_MS": 120000,
  "AGENT_RUNTIME_MAX_STEPS": 6,
  "MASTRA_AGENT_MODEL": "",
  "AGENT_MCP_ENABLED": false,
  "AGENT_MCP_SERVERS_JSON": {},
  "AGENT_MCP_TIMEOUT_MS": 15000,
  "LLM_PARSE_INTENT_TIMEOUT_MS": 12000,
  "LLM_GENERATE_REPLY_TIMEOUT_MS": 30000,
  "LLM_STREAM_STALL_TIMEOUT_MS": 12000,
  "AUTO_ROUTER_DISCOVERY_BREAKER_FAILURES": 3,
  "AUTO_ROUTER_DISCOVERY_BREAKER_COOLDOWN_MS": 45000,
  "ATTACHMENT_SCAN_MODE": "off",
  "ATTACHMENT_SCAN_FAIL_CLOSED": false,
  "ATTACHMENT_SCAN_TIMEOUT_MS": 5000,
  "METRICS_PUSHGATEWAY_URL": "",
  "METRICS_JOB_NAME": "autopilot-backend",
  "METRICS_INSTANCE_ID": "",
  "METRICS_PUSH_INTERVAL_MS": 15000,
  "METRICS_PUSH_TIMEOUT_MS": 5000,
  "METRICS_SNAPSHOT_PATH": ""
}

📖 Usage Guide

1. Initial Setup

Access the UI at http://localhost:3000 and complete the setup:

  1. Configure LLM Provider - Go to Settings → AI Providers

    • Add your OpenAI, Gemini, or Ollama credentials
    • Set a default provider

  2. Connect Workflow Providers - Settings → Workflow Providers

    • Add n8n webhook URL
    • Configure Zapier/Make credentials
    • Set up custom endpoints

  3. Configure Authentication - Enable Google OAuth or use local auth

2. Creating Workflows

Navigate to Workflows → New Workflow:

Key: daily_report
Name: Daily Analytics Report
Provider: n8n
Execution Endpoint: https://n8n.example.com/webhook/daily-report
Method: POST
Requires Approval: true  # Enable approval gates

3. Chat Interaction

In the Chat interface:

User: Run the daily report for yesterday
Assistant: I'll trigger the daily report workflow for you.
       [Workflow triggered: daily_report]
       Status: waiting_approval

User: Approve it
Assistant: Approved and running... ✓
       Workflow completed successfully.
       [View Results]

4. Webhook Integration

AutoPilot receives workflow completion callbacks:

n8n Webhook Node:

// HTTP Request node configuration
Method: POST
URL: https://autopilot.example.com/api/webhooks/n8n
Headers: {
  "x-webhook-secret": "your-secret"
}
Body: {
  "type": "completed",
  "runId": "{{ $runId }}",
  "result": {{ $json }}
}

Universal Callback:

POST /api/webhooks/callback
Headers: x-webhook-secret: <secret>
Body: {
  "traceId": "trace_123",
  "workflowKey": "daily_report",
  "provider": "n8n",
  "status": "completed",
  "result": { ... }
}

🏗️ Architecture

┌─────────────────────────────────────────────────────────────┐
│                        Browser (SolidJS)                    │
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────────────┐   │
│  │  Chat UI    │  │ Workflows   │  │ Settings/Providers  │   │
│  └──────┬──────┘  └──────┬──────┘  └──────────┬──────────┘   │
└─────────┼────────────────┼────────────────────┼──────────────┘
          │                │                    │
          └────────────────┴────────────────────┘
                           │
                    ┌──────▼──────┐
                    │   Express   │
                    │   Backend   │
                    │   (Bun)     │
                    └──────┬──────┘
           ┌───────────────┼───────────────┐
           │               │               │
    ┌──────▼──────┐ ┌──────▼──────┐ ┌──────▼──────┐
    │ PostgreSQL  │ │  Workflow   │ │    LLM      │
    │  (Drizzle)  │ │  Providers  │ │  Providers  │
    └─────────────┘ └─────────────┘ └─────────────┘

Runtime Model

  • Primary Runtime: AgentService (Mastra-based agent runtime)
  • Fallback Runtime: OrchestratorService (deterministic orchestration)
  • Chat routes attempt agent runtime first; fallback to orchestrator on failure

🛠️ Development

Local Setup

# Clone and install
git clone <repo>
cd autopilot
bun install

# Environment
cp .env.example .env
# Edit .env with your credentials

# Database (from apps/backend)
bun run db:preflight
bun run db:migrate

# Run all services
bun run dev

Default URLs:

Build

# Build all workspaces
bun run build

# Type check
bun run typecheck

# Run tests
bun run test

🔧 Troubleshooting

Container Won't Start

# Check logs
docker logs autopilot

# Verify environment variables
docker run --rm ajayduddi/autopilot:latest env | grep -E "(DATABASE_URL|AUTH_COOKIE_SECRET)"

Database Connection Issues

# Test connectivity
docker exec -it autopilot ping your-db-host

# Run preflight check
bun run db:preflight

Webhook Callbacks Not Received

  1. Verify CALLBACK_BASE_URL points to public URL
  2. Check webhook secret configuration
  3. Ensure firewall allows incoming POST to /api/webhooks/*

LLM Provider Errors

  • Check provider credentials in Settings → AI Providers
  • Verify API key encryption key is set (PROVIDER_API_KEY_ENCRYPTION_KEY)
  • For Ollama, ensure OLLAMA_URL is accessible from container

📚 Documentation

📄 License

Apache License 2.0 - See LICENSE for details.

Built with 🦋 Bun · ⚡ SolidJS · 🐘 PostgreSQL

About

AutoPilot transforms how you interact with automation workflows. Instead of manual triggers and complex dashboards, simply chat with your automations. It connects to multiple workflow providers and LLMs, giving you a unified conversational interface for execution, monitoring, and approval workflows, optimized for production reliability and scale

Topics

automation ai zapier autopilot make chat-application workflows sim n8n chatplatform ai-agent orchrestration work-with-llms

Resources

Readme

License

Apache-2.0 license
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages