Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
19 changes: 19 additions & 0 deletions FlowVision/FlowVision.csproj
Original file line number Diff line number Diff line change
Expand Up @@ -168,6 +168,10 @@
<Reference Include="System.Threading.Tasks.Extensions, Version=4.2.4.0, Culture=neutral, PublicKeyToken=cc7b13ffcd2ddd51, processorArchitecture=MSIL">
<HintPath>..\packages\System.Threading.Tasks.Extensions.4.6.3\lib\net462\System.Threading.Tasks.Extensions.dll</HintPath>
</Reference>
<Reference Include="Tesseract">
<HintPath>..\packages\Tesseract.5.2.0\lib\net48\Tesseract.dll</HintPath>
<Private>True</Private>
</Reference>
<Reference Include="Microsoft.ML.OnnxRuntime">
<HintPath>..\packages\Microsoft.ML.OnnxRuntime.Managed.1.21.1\lib\netstandard2.0\Microsoft.ML.OnnxRuntime.dll</HintPath>
<Private>True</Private>
Expand Down Expand Up @@ -316,7 +320,22 @@
</ItemGroup>
<Copy SourceFiles="@(OnnxRuntimeNative)" DestinationFolder="$(OutputPath)" SkipUnchangedFiles="true" />
</Target>
<!-- Copy Tesseract native DLLs to output directory -->
<Target Name="CopyTesseractNative" AfterTargets="AfterBuild">
<ItemGroup>
<TesseractNative Include="..\packages\Tesseract.5.2.0\x64\*.dll" />
</ItemGroup>
<Copy SourceFiles="@(TesseractNative)" DestinationFolder="$(OutputPath)" SkipUnchangedFiles="true" />
</Target>
<!-- Copy OmniParser model to output directory -->
<Target Name="CopyOmniParserModel" AfterTargets="AfterBuild">
<ItemGroup>
<OmniParserModel Include="Models\icon_detect.onnx" />
</ItemGroup>
<Copy SourceFiles="@(OmniParserModel)" DestinationFolder="$(OutputPath)\models" SkipUnchangedFiles="true" />
</Target>
<Import Project="..\packages\CefSharp.Common.135.0.170\build\CefSharp.Common.targets" Condition="Exists('..\packages\CefSharp.Common.135.0.170\build\CefSharp.Common.targets')" />
<Import Project="..\packages\Tesseract.5.2.0\build\Tesseract.targets" Condition="Exists('..\packages\Tesseract.5.2.0\build\Tesseract.targets')" />
<Import Project="..\packages\Fody.6.9.2\build\Fody.targets" Condition="Exists('..\packages\Fody.6.9.2\build\Fody.targets')" />
<Import Project="..\packages\Costura.Fody.6.0.0\build\Costura.Fody.targets" Condition="Exists('..\packages\Costura.Fody.6.0.0\build\Costura.Fody.targets')" />
<Import Project="..\packages\Microsoft.Playwright.1.52.0\build\Microsoft.Playwright.targets" Condition="Exists('..\packages\Microsoft.Playwright.1.52.0\build\Microsoft.Playwright.targets')" />
Expand Down
13 changes: 13 additions & 0 deletions FlowVision/lib/Classes/ai/LMStudioActioner.cs
Original file line number Diff line number Diff line change
Expand Up @@ -185,11 +185,24 @@ public async Task<string> ExecuteAction(string actionPrompt)

var response = responseBuilder.ToString();

// Log the response for debugging
PluginLogger.LogInfo("LMStudioActioner", "ExecuteAction",
$"AI Response: {(string.IsNullOrEmpty(response) ? "(empty)" : $"{response.Length} chars")}");

// Add assistant response to history
if (!string.IsNullOrEmpty(response))
{
actionerHistory.Add(new ChatMessage(ChatRole.Assistant, response));
}
else
{
// If response is empty, check if there were tool calls and provide a default response
var defaultResponse = "I executed the requested action successfully.";
PluginLogger.LogInfo("LMStudioActioner", "ExecuteAction",
"Empty response from model, using default response");
actionerHistory.Add(new ChatMessage(ChatRole.Assistant, defaultResponse));
return defaultResponse;
}

return response;
}
Expand Down
105 changes: 49 additions & 56 deletions docs/Getting-Started.md
Original file line number Diff line number Diff line change
Expand Up @@ -3,94 +3,87 @@ layout: default
title: Getting Started
---

# Getting Started with Recursive Control
# First Flight Checklist

This guide will help you get up and running with Recursive Control quickly.
Strap into Mission Control and run your first full-loop automation in under ten minutes.

## Your First Commands
## Cue the Agents

Once installed and configured, you can start using natural language commands:
Once the cockpit is configured you can talk in plain-text, mission-style briefs.

### Basic Examples
### Launch Scripts

1. **File Management**
1. **File Ops Circuit**
```
"Open File Explorer and navigate to my Documents folder"
"Create a new folder called Projects on my Desktop"
"Rename all .txt files in this folder to .md"
Open File Explorer and navigate to my Documents folder.”
Create a new folder called Projects on my Desktop.”
Rename all .txt files in this folder to .md.”
```

2. **Application Control**
2. **App Relay**
```
"Open Notepad and type Hello World"
"Launch Chrome and navigate to github.com"
"Open Excel and create a new spreadsheet"
Open Notepad and type Hello World.”
Launch Chrome and navigate to github.com.”
Open Excel and create a new spreadsheet.”
```

3. **Screen Capture**
3. **Vision Probe**
```
"Take a screenshot of the current window"
"Capture the entire screen"
Take a screenshot of the current window.”
Capture the entire screen.”
```

4. **Web Automation**
4. **Browser Drone**
```
"Open a browser and search for AI automation tools"
"Fill out this form with my information"
"Extract data from this webpage"
Open a browser and search for AI automation tools.”
Fill out this form with my information.”
Extract data from this webpage.”
```

## Understanding the Multi-Agent System
## Meet the Crew

Recursive Control uses a sophisticated 3-agent architecture:
- **Hermes (Coordinator)** — interprets intent, sets the heading
- **Daedalus (Planner)** — drafts the maneuver plan step by step
- **Talos (Executor)** — performs every keystroke, click, and command

1. **Coordinator Agent**: Understands your request and determines the best approach
2. **Planner Agent**: Creates a step-by-step plan to accomplish the task
3. **Executor Agent**: Executes the plan using available plugins
Agents talk to you in real-time so you can abort, adjust, or cheer as they work.

This system ensures tasks are completed efficiently and accurately.
## Mission Briefing Tips

## Tips for Better Results
- **Paint the picture** → “Move all PDF files from Downloads to Documents” beats “Do something with files.”
- **Chunk the mission** → Run multi-stage workflows in segments and confirm each leg.
- **Talk human** → Natural language with intent and context works best; no scripting required.

### Be Specific
❌ "Do something with files"
✅ "Move all PDF files from Downloads to Documents folder"
## Plugin Bay at a Glance

### Break Down Complex Tasks
For very complex workflows, break them into smaller steps and verify each step.
- Keyboard + Mouse autopilot
- Command line (CMD + PowerShell)
- Screen capture + annotation
- Window switching + focus
- Playwright browser automation
- Remote control HTTP bridge

### Use Natural Language
You don't need to use technical commands - just describe what you want in plain English.
## Next Flight Plans

## Available Plugins
- Tour the [UI Features Atlas](UI-Features.html)
- Study the [Multi-Agent Architecture](Multi-Agent-Architecture.html)
- Customize personas via [System Prompts Reference](System-Prompts-Reference.html)
- Join the [Discord flight crew](https://discord.gg/mQWsWeHsVU)

Your commands can utilize these built-in plugins:
## First-Time FAQs

- **Keyboard & Mouse**: Automate input and clicks
- **Command Line**: Execute CMD and PowerShell commands
- **Screen Capture**: Take and analyze screenshots
- **Window Management**: Control application windows
- **Web Browser**: Automate websites with Playwright
- **Remote Control**: Accept commands via HTTP API
**How do I know it’s engaged?**

## Next Steps
Watch the live reasoning feed and action timeline.

- Explore the [UI Features](UI-Features.html) guide
- Learn about [Multi-Agent Architecture](Multi-Agent-Architecture.html)
- Check out [System Prompts](System-Prompts-Reference.html) for customization
- Join our [Discord](https://discord.gg/mQWsWeHsVU) for community support
**Something went sideways.**

## Common First-Time Questions
Hit the abort control, then consult the [Troubleshooting Console](Troubleshooting.html).

**Q: How do I know if it's working?**
A: The UI will show the agent's thinking process and actions in real-time.
**Can I reshape its behavior?**

**Q: What if something goes wrong?**
A: You can interrupt execution at any time. Check the [Troubleshooting](Troubleshooting.html) guide for help.

**Q: Can I customize the behavior?**
A: Yes! You can adjust system prompts, enable/disable plugins, and configure various settings.
Absolutely. Tweak prompts, toggle plugins, and save profiles.

---

Ready to dive deeper? Check out our [advanced documentation](Multi-Agent-Architecture.html) or join the [community](https://discord.gg/mQWsWeHsVU)!
Ready for more? Dive into [advanced documentation](Multi-Agent-Architecture.html) or sync with the [community](https://discord.gg/mQWsWeHsVU).
63 changes: 31 additions & 32 deletions docs/Installation.md
Original file line number Diff line number Diff line change
Expand Up @@ -3,46 +3,45 @@ layout: default
title: Installation
---

# Installation Guide
# Installation Capsule

Get Recursive Control up and running on your Windows system.
Your mission dock for landing Recursive Control on Windows. Choose your payload, confirm dependencies, and ignite the thrusters.

## System Requirements
## Flight Readiness Checklist

### Minimum Requirements
### Baseline Specs
- **Operating System**: Windows 10 or Windows 11
- **.NET Framework**: 4.8 or later
- **RAM**: 4 GB minimum (8 GB recommended)
- **Disk Space**: 500 MB for application files

### Additional Requirements
### Network & Auth
- Internet connection for AI model API access
- API key for your preferred AI provider (OpenAI, Azure OpenAI, Anthropic, Google, etc.)

## Installation Steps
## Deploy Options

### Option 1: Download Pre-built Release (Recommended)
### Rapid Drop (Recommended)

1. **Download the Latest Release**
- Visit the [Releases page](https://github.com/flowdevs-io/Recursive-Control/releases)
- Download the latest `recursivecontrol.zip` or installer
- Extract to your preferred location
1. **Pull the latest build**
- Hop to the [Releases hangar](https://github.com/flowdevs-io/Recursive-Control/releases)
- Snag the newest `recursivecontrol.zip` payload
- Extract anywhere friendly (no admin elevation required)

2. **Run the Application**
- Double-click `recursivecontrol.exe`
- The application will launch and prompt for initial setup
2. **Boot the cockpit**
- Launch `recursivecontrol.exe`
- First boot unlocks the setup co-pilot

3. **Configure Your AI Provider**
- Click the settings/configuration button
- Select your AI provider (OpenAI, Azure, Anthropic, etc.)
- Enter your API key
- Choose your preferred model
3. **Wire your pilot**
- Open Settings → Providers
- Pick your model stack (OpenAI, Azure, Claude, Gemini, LM Studio…)
- Drop in your API credentials

4. **Test the Installation**
- Try a simple command like "What can you do?"
- Verify the AI responds correctly
4. **Ping the tower**
- Ask "What can you do?"
- Confirm you see agent chatter + completion

### Option 2: Build from Source
### Source Build (Engineers’ Track)

For developers who want to build from source:

Expand All @@ -63,9 +62,9 @@ dotnet build
dotnet run --project FlowVision
```

## Initial Configuration
## Configure Your Pilot

### Setting Up Your AI Provider
### OpenAI

#### OpenAI
1. Get your API key from [OpenAI Platform](https://platform.openai.com)
Expand Down Expand Up @@ -97,7 +96,7 @@ dotnet run --project FlowVision
- Provider: LM Studio
- Endpoint: http://localhost:1234 (or your configured port)

### Plugin Configuration
### Plugin Bay

Enable or disable plugins based on your needs:

Expand All @@ -110,7 +109,7 @@ Enable or disable plugins based on your needs:
- ⚠️ Playwright (requires additional setup)
- ⚠️ Remote Control (enable for HTTP API)

## Verifying Installation
## System Checks

Run these test commands to verify everything works:

Expand All @@ -121,17 +120,17 @@ Run these test commands to verify everything works:

If all tests pass, you're ready to go!

## Troubleshooting Installation Issues
## Troubleshooting Bay

### .NET Framework Not Found
### .NET Framework Missing
- Download and install [.NET Framework 4.8](https://dotnet.microsoft.com/download/dotnet-framework/net48)

### API Key Errors
- Verify your API key is correct
- Check that your API provider account has credits/active subscription
- Ensure internet connection is working

### Application Won't Start
### Application Wont Start
- Run as Administrator
- Check Windows Event Viewer for error details
- Verify all dependencies are installed
Expand All @@ -141,13 +140,13 @@ If all tests pass, you're ready to go!
- Increase available RAM
- Consider using a lighter AI model

## Next Steps
## Next Flight Plans

- Continue to [Getting Started](Getting-Started.html) guide
- Explore [UI Features](UI-Features.html)
- Join our [Discord community](https://discord.gg/mQWsWeHsVU)

## Uninstallation
## Uninstall Playbook

To remove Recursive Control:
1. Delete the application folder
Expand Down
8 changes: 6 additions & 2 deletions docs/_config.yml
Original file line number Diff line number Diff line change
Expand Up @@ -6,15 +6,19 @@ description: AI-Powered Computer Control for Windows
baseurl: "/Recursive-Control"
url: "https://flowdevs-io.github.io"

# Theme (GitHub Pages supported themes)
remote_theme: pages-themes/cayman@v0.2.0
theme: null

# Plugins (GitHub Pages whitelist)
plugins:
- jekyll-remote-theme
- jekyll-seo-tag
- jekyll-sitemap

# Sass pipeline
sass:
sass_dir: _sass
style: compressed

# Markdown settings
markdown: kramdown
kramdown:
Expand Down
Loading