Unleash AI Automation: A Deep Dive into Manifold

Tired of repetitive tasks? Want to harness the power of AI for your workflows? Manifold is your answer. This powerful platform lets you automate complex processes using cutting-edge AI models, streamline your operations, and boost productivity. While still under active development, Manifold offers a glimpse into the future of AI-driven automation.

What is Manifold and Why Should You Care?

Manifold is a workflow automation platform fueled by AI. It integrates seamlessly with popular AI endpoints like OpenAI, Google Gemini, and more. This allows you to automate a wide array of tasks, from text and image generation to advanced semantic search, all in one place.

Key Benefits of Using Manifold:

• Automate Repetitive Tasks: Free up your time by automating workflows with AI.
• Integrate with Leading AI Models: Leverage the power of OpenAI, Google Gemini, and others.
• Semantic Search: Find information quickly and accurately using PGVector and SEFII.
• Text & Image Generation: Generate content automatically.
• Highly Customizable: Adapts to various workflows.

Getting Started: Quick Start Guide

Ready to jump in? The easiest way to experience Manifold is with pre-built binaries. Here's how:

1. Download the Binary: Grab the right version for your OS from the releases page.
   • macOS: manifold-darwin-universal.zip
   • Linux: manifold-linux-amd64.zip or manifold-linux-arm64.zip
2. Extract and Navigate: Unzip the file and enter the extracted directory via your terminal.
3. Configuration: Create config.yaml in the same directory. Use config.yaml.example as your template.
4. Run Manifold:

   chmod +x manifold-*
   ./manifold-*
   

   Copy

Installation from Source: A Step-by-Step Guide

Want more control? Installing from source is the way to go. Here's the process:

1. Clone the Repository:

   git clone https://github.com/intelligencedev/manifold.git
   cd manifold
   

   Copy
2. Initialize Submodules: Fetch necessary dependencies like llama.cpp and pgvector.

   git submodule update --init --recursive
   

   Copy
3. Choose an Image Generation Backend: Pick either ComfyUI (cross-platform) or MFlux (M-series Macs only) and follow their respective installation guides.
4. Configure Manifold: Create a config.yaml file (using config.yaml.example) and place it in the same path as your main.go file for development, or next to your Manifold binary. Your config.yaml file dictates the behavior of Manifold and is critical for defining database connection strings, defining your API keys, and more.
5. Build and Run:

   cd frontend
   nvm use 20
   npm install
   npm run build
   cd ..
   go build -ldflags="-s -w" -trimpath -o ./dist/manifold
   cd dist
   ./manifold
   

   Copy

Configuration: Unleashing Manifold's Potential with config.yaml

The config.yaml file is your control panel. Here's an example:

host: localhost
port: 8080
data_path: ./data
jaeger_host: localhost:6831 # Optional Jaeger tracing

# API Keys (optional integrations)
anthropic_key: "..."
openai_api_key: "..."
google_gemini_key: "..."
hf_token: "..."

# Database Configuration
database:
  connection_string: "postgres://myuser:changeme@localhost:5432/manifold"

# Completion and Embedding Services
completions:
  default_host: "http://localhost:32186/v1/chat/completions" # Default: local llama-server
  api_key: ""
embeddings:
  host: "http://localhost:32184/v1/embeddings" # Default: local llama-server
  api_key: ""
  dimensions: 768
  embed_prefix: "search_document: "
  search_prefix: "search_query: "
reranker:
  host: "http://localhost:32185/v1/rerank" # Default: local llama-server

• Database Credentials: Update myuser and changeme to match your PGVector setup.
• Model Server: Adjust default_host and embeddings.host to your chosen model server.

Accessing Manifold: Your Gateway to AI Automation

Open your browser and head to http://localhost:8080. If you changed the port in config.yaml, use that port instead.

Troubleshooting: Conquering Common Hurdles

Encountering issues? Here's a quick guide:

• Port Conflict: Change the port in config.yaml or terminate conflicting processes.
• PGVector Connectivity: Check your database.connection_string.
• Missing Config File: Make sure config.yaml is in the correct directory.

Diving Deeper: Running Manifold in Development Mode

For development, ensure your config.yaml is at the project root, then:

cd frontend
nvm use 20
npm install
npm run dev

In another terminal, start the Go backend.

Semantic Search with PGVector: Find What You Need, Fast

Manifold uses PGVector and the SEFII engine to offer robust semantic search capabilities. This means you can search for information based on meaning, not just keywords. The semantic search implementation in Manifold paired with clever indexing significantly improves retrieval times.

Contributing: Shape the Future of Manifold

Want to help build the future of AI automation? Manifold welcomes contributions. Check out the open issues and submit pull requests.

Conclusion: Embrace the Power of Manifold

Manifold offers a powerful and flexible platform for AI-driven workflow automation. While still under development, its potential is undeniable. By following this guide, you can get started with Manifold and begin unlocking the power of AI for your own projects.