FlowFRAM Connector

About

FlowFRAM Connector bridges the FlowFRAM web application with your local services, enabling two powerful capabilities from a single extension:

▶ RUNTIME PROXY — Run high-performance FRAM (Functional Resonance Analysis Method) simulations on your own machine or private network using FlowFRAM's distributed runtime agent.

▶ OLLAMA PROXY — Use your local Ollama LLM instance for AI-assisted analysis directly from flowfram.com. No API keys needed — your models, your hardware, your data.

🚀 Key Features:
- Dual-purpose: Runtime + Ollama proxy in a single extension
- Seamless connection between FlowFRAM and local runtime agents
- Local LLM inference via Ollama (generate, chat, list models)
- Real-time flow deployment and execution monitoring via SSE
- Tabbed configuration: separate settings for Runtime and Ollama
- Visual badge indicator: R (Runtime), O (Ollama), R·O (both), or red ! (error)
- Automatic connection status detection
- Origin header stripping for full Ollama compatibility
- On-demand host permissions — only requests access when needed
- Secure local communication without exposing data to external servers

💡 Why use a local runtime?
- Execute thousands of simulation iterations per second
- Access local APIs, databases, and network resources
- Keep sensitive data on your own infrastructure
- Work offline once connected

🧠 Why use local Ollama?
- Run AI analysis with your own hardware — no cloud API costs
- Full privacy: your prompts and data never leave your machine
- Use any Ollama-supported model (Llama 3, Mistral, Gemma, Phi, etc.)
- No API keys or subscriptions required

🔧 How it works:
1. Install the FlowFRAM Runtime Agent on your machine (Docker Hub image available) and/or install Ollama with a model
2. Install this extension and configure the service URLs (defaults: localhost:3010 for Runtime, localhost:11434 for Ollama)
3. Open FlowFRAM — the extension automatically bridges your browser to local services, bypassing CORS restrictions

📋 Requirements:
- For Runtime: FlowFRAM Runtime Agent running locally or on your network (Docker: cgoudouris/flowfram-runtime)
- For Ollama: Ollama installed with at least one model pulled (e.g., ollama pull llama3)
- Both services are optional — enable only what you need

🔒 Privacy:
- No data collection, no analytics, no tracking
- Configuration stored locally via chrome.storage.local
- All communication stays between your browser and your local services

This extension is part of PhD research on complex systems modeling using the FRAM methodology.
Learn more at https://flowfram.com/agent