Skip to content

patrickkarle/ted_skinner_project

BranchesTags

Repository files navigation

FullIntel Agent

A powerful desktop application for AI-powered research workflows and intelligent conversations. Execute multi-phase research protocols, engage in LLM chat, and generate comprehensive executive briefs on companies, industries, and emerging technologies.

Version Platform

Features

🤖 Multi-Provider AI Support

  • 4 Built-in Providers - Anthropic Claude, OpenAI GPT, Google Gemini, DeepSeek
  • Custom Provider Support - Add your own API endpoints (e.g., local Ollama, custom OpenAI-compatible APIs)
  • Per-Provider API Keys - Configure different keys for each provider
  • Model Selection - Choose from 12+ models across providers

💬 Dual-Mode Operation

  • Chat Mode - Direct LLM conversations without requiring research workflows
  • Research Mode - Execute structured multi-phase research protocols
  • Seamless Switching - Toggle between modes with a single click
  • Post-Research Followup - Continue conversations after research completes

📋 Manifest System

  • 7 Pre-configured Protocols - Ready-to-use research workflows for different analysis types
  • Manifest Editor - Full-featured YAML editor with syntax validation
  • Create Custom Manifests - Design your own research workflows from scratch
  • Load External Manifests - Import YAML files from anywhere on your system
  • Live Phase Preview - See validated phases before running
  • Manifest Management - Rename, remove, and organize your manifest library

📊 Research Session Management

  • Full Session Persistence - All sessions saved to local SQLite database
  • Session History - Browse all past research sessions with search
  • Session Resume - Continue incomplete research from where you left off
  • Phase-Level Storage - Each phase output saved individually with timestamps
  • Project Organization - Group related sessions into named projects
  • Archive System - Archive and restore both projects and sessions

✏️ Prompt Editing & Relaunch

  • View Prompts - See exactly what system/user prompts were sent to the LLM
  • Edit Prompts - Modify prompts before re-running a phase
  • Phase Relaunch - Re-run any individual phase with original or modified prompts
  • Prompt History - All prompts preserved for reproducibility

🔄 Real-Time Features

  • Streaming Output - Watch AI responses appear token-by-token
  • Live Phase Status - Visual progress indicators for each phase
  • Activity Logs - Detailed execution logs with timing information
  • Abort Support - Cancel running operations at any time

📤 Export & Output

  • Copy to Clipboard - One-click copy of full research output
  • Markdown Export - Save results as formatted Markdown files
  • Print Support - Print-optimized views for physical copies
  • Per-Phase Export - Export individual phase outputs separately

🔐 Security & Privacy

  • Local-First Design - All data stored locally on your machine
  • Encrypted Credentials - API keys and passwords securely stored
  • User Accounts - Multi-user support with separate data isolation
  • No External Telemetry - Only API calls to your selected LLM provider

Installation

Available Downloads

All binaries are available on the GitHub Releases page:

Platform File Type
Windows fullintel-agent_0.1.1_x64-setup.exe NSIS installer (recommended)
Windows fullintel-agent_0.1.1_x64_en-US.msi MSI installer (enterprise)
macOS Intel fullintel-agent_0.1.1_x64.dmg Disk image
macOS Apple Silicon fullintel-agent_0.1.1_aarch64.dmg Disk image (M1/M2/M3)
Linux fullintel-agent_0.1.1_amd64.AppImage Universal (any distro)
Linux fullintel-agent_0.1.1_amd64.deb Debian/Ubuntu package

Windows

  1. Download fullintel-agent_0.1.1_x64-setup.exe from Releases
  2. Run the installer and follow the wizard
  3. Launch "FullIntel Agent" from the Start Menu

Alternative: Download the .msi installer for enterprise deployment.

macOS

Intel Macs (x64)

  1. Download fullintel-agent_0.1.1_x64.dmg from Releases
  2. Open the DMG and drag the app to Applications
  3. Launch from Applications folder

Apple Silicon (M1/M2/M3)

  1. Download fullintel-agent_0.1.1_aarch64.dmg from Releases
  2. Open the DMG and drag the app to Applications
  3. Launch from Applications folder

Note: On first launch, you may need to right-click → Open to bypass Gatekeeper since the app is not notarized.

Linux

AppImage (Universal)

  1. Download fullintel-agent_0.1.1_amd64.AppImage from Releases
  2. Make it executable: chmod +x fullintel-agent_*.AppImage
  3. Run it: ./fullintel-agent_*.AppImage

Debian/Ubuntu (.deb)

  1. Download fullintel-agent_0.1.1_amd64.deb from Releases
  2. Install: sudo dpkg -i fullintel-agent_*.deb
  3. Launch from your application menu or run fullintel-agent

System Requirements

Platform Requirements
Windows Windows 10/11 (64-bit), WebView2 Runtime
macOS macOS 10.15 (Catalina) or later
Linux GTK 3, WebKitGTK 4.1, glibc 2.31+

API Key Requirement

You'll need an API key from at least one supported provider:

Quick Start

First Time Setup

  1. Create Account - Set up a local user account on first launch
  2. Configure API Key - Go to Settings (gear icon) and enter your API key for any supported provider

Chat Mode (Default)

  1. Select Model - Choose your AI provider and model from the dropdowns
  2. Type Message - Enter your question or prompt in the input field
  3. Send - Press Enter or click Send to chat with the AI
  4. Continue Conversation - Follow up with additional questions

Research Mode

  1. Switch Mode - Click "Research" toggle (next to Chat)
  2. Select Manifest - Choose a research protocol from the dropdown
  3. Enter Topic - Type a company name or research subject
  4. Run Research - Click "Run" and watch results stream in phase by phase
  5. Follow Up - Ask questions about the research after it completes

Session Management

  1. View History - Browse past sessions in the left sidebar
  2. Resume Session - Click any incomplete session to resume from last phase
  3. Edit & Relaunch - View prompts for any phase, edit them, and re-run
  4. Organize - Create projects to group related research sessions

Research Manifests

The application includes 7 pre-configured research protocols:

Manifest Purpose
Fullintel Process General company research and competitive analysis
Deep Dive Industry Comprehensive industry sector analysis
Frontier Tech Emerging technology and innovation research
Inductive Company Bottom-up company analysis from public signals
Intellectual Capital Talent mapping and human capital analysis
Value Chain Supply chain and value network analysis
Venture Momentum Startup and investment landscape research

Manifest Structure

Each manifest defines a multi-phase research workflow in YAML format:

manifest:
  id: "Protocol-ID"
  version: "1.0.0"
  name: "Protocol Name"
  description: "What this protocol analyzes"

phases:
  - id: "PHASE-01"
    name: "Phase Name"
    instructions: |
      Detailed instructions for the AI researcher...
    output_schema: "SchemaName"

  - id: "PHASE-02"
    name: "Analysis Phase"
    dependencies: ["PHASE-01"]
    instructions: |
      Build on previous phase results...

quality_gates:
  - phase: "PHASE-01"
    check: "Validation criteria"
    fail_action: "RETRY"

Creating Custom Manifests

Using the Manifest Editor (Recommended)

  1. Click the "+ New" button in the Manifests section of the sidebar
  2. The Manifest Editor opens with a template
  3. Edit the YAML content with your custom phases and instructions
  4. Click "Validate" to check for errors
  5. Click "Save" to save your manifest
  6. Click "Use Manifest" to load it immediately

Editing Existing Manifests

  1. Click the "Edit" (pencil icon) next to any manifest
  2. Modify the YAML content in the editor
  3. Validate and save your changes
  4. Changes are immediately available for use

Loading External Files

  1. Click "Open" in the Manifest Editor toolbar
  2. Select any .yaml or .yml file from your system
  3. Validate the content
  4. Save to add it to your manifest library

Manual File Creation

You can also create manifests manually:

  1. Create a new .yaml file in the manifests/ directory
  2. Follow the manifest schema structure
  3. Restart the app to see it in the dropdown

Architecture

Technology Stack

  • Frontend: React 18 + TypeScript + Vite
  • Backend: Rust (Tauri v2)
  • Database: SQLite with rusqlite
  • AI Integration: Multi-provider support with streaming
    • Anthropic Claude (Claude Opus 4, Sonnet 4.5, Haiku 3.5)
    • OpenAI GPT (GPT-4o, GPT-4 Turbo, GPT-3.5 Turbo)
    • Google Gemini (Gemini 1.5 Pro, Gemini 1.5 Flash)
    • DeepSeek (DeepSeek Chat, DeepSeek Coder)
  • Styling: CSS with design tokens

Project Structure

ted_skinner_project/
├── src/                    # React frontend
│   ├── App.tsx            # Main application component
│   ├── App.css            # Application styles
│   └── components/        # UI components
├── src-tauri/             # Rust backend
│   ├── src/
│   │   ├── main.rs       # Tauri commands & app setup
│   │   ├── agent.rs      # Research workflow engine
│   │   ├── auth.rs       # Authentication & database
│   │   ├── llm.rs        # LLM API integration
│   │   └── manifest.rs   # Manifest parsing
│   └── tauri.conf.json   # Tauri configuration
├── manifests/             # Research protocol definitions
│   ├── fullintel_process_manifest.yaml
│   ├── Deep-Dive-Industry-Protocol.yaml
│   ├── frontier-tech-protocol.yaml
│   ├── Inductive-Company-Protocol.yaml
│   ├── Intellectual-capital-protocol.yaml
│   ├── value-chain-protocol.yaml
│   └── venture-momentum-protocol.yaml
└── docs/                  # Documentation

Development

Prerequisites

Platform-Specific Dependencies

Windows:

  • WebView2 (usually pre-installed on Windows 10/11)
  • Visual Studio Build Tools with C++ workload

macOS:

  • Xcode Command Line Tools: xcode-select --install

Linux (Debian/Ubuntu):

sudo apt install libgtk-3-dev libwebkit2gtk-4.1-dev libayatana-appindicator3-dev librsvg2-dev

Setup

# Install dependencies
npm install

# Run in development mode
npm run tauri dev

# Build for production (current platform)
npm run tauri build

# Build for specific target
npm run tauri build -- --target x86_64-apple-darwin    # macOS Intel
npm run tauri build -- --target aarch64-apple-darwin   # macOS Apple Silicon
npm run tauri build -- --target x86_64-unknown-linux-gnu  # Linux
npm run tauri build -- --target x86_64-pc-windows-msvc    # Windows

Testing

# Run all Rust tests
cd src-tauri && cargo test

# Run specific test battery
cargo test battery1  # Unit tests
cargo test battery2  # Integration tests
cargo test battery3  # System tests

Configuration

API Key Setup

Get an API key from any supported provider:

Provider Console URL Notes
Anthropic console.anthropic.com Claude models
OpenAI platform.openai.com GPT models
Google aistudio.google.com Gemini models
DeepSeek platform.deepseek.com Cost-effective alternative

Then in the app:

  1. Click Settings (gear icon)
  2. Select your provider from the dropdown
  3. Enter your API key
  4. Click Save

Model Selection

Choose models based on your needs:

Anthropic Claude

  • Claude Opus 4 - Highest quality, complex analysis
  • Claude Sonnet 4.5 - Recommended balance of speed/quality
  • Claude Haiku 3.5 - Fastest, lower cost

OpenAI GPT

  • GPT-4o - Latest multimodal model
  • GPT-4 Turbo - High capability, faster than GPT-4
  • GPT-3.5 Turbo - Fast and economical
  • o1-preview - Advanced reasoning model
  • o3-mini - Efficient reasoning model

Google Gemini

  • Gemini 1.5 Pro - Advanced reasoning
  • Gemini 1.5 Flash - Fast responses

DeepSeek

  • DeepSeek Chat - General purpose
  • DeepSeek Coder - Code-focused analysis

Custom Providers

Add your own AI providers for additional flexibility:

  1. Open Settings - Click the gear icon
  2. Go to Custom Providers - Select the "Custom Providers" tab
  3. Add Provider - Enter:
    • Name: Display name for the provider
    • Base URL: API endpoint (e.g., http://localhost:11434/v1)
    • Model ID: Model identifier (e.g., llama3.2)
    • API Key: Optional, depending on provider
  4. Save - Your custom provider appears in the model dropdown

Common Use Cases:

  • Local Ollama instances
  • OpenAI-compatible APIs (Together AI, Groq, etc.)
  • Self-hosted models
  • Enterprise private deployments

Data Storage

All user data is stored locally on your machine:

Platform Data Directory
Windows %APPDATA%\com.fullintel.agent\
macOS ~/Library/Application Support/com.fullintel.agent/
Linux ~/.local/share/com.fullintel.agent/
File Purpose
users.db User database and research sessions
config.json Application configuration

No data is sent to external servers except for the AI API calls to your selected provider.

Usage Guide

Chat Mode

The application defaults to Chat Mode for quick AI conversations:

  1. Start Chatting - Type your message in the input field and press Enter
  2. View Response - AI responses stream in real-time
  3. Continue Conversation - Ask follow-up questions to build context
  4. Switch Models - Change AI models mid-conversation if needed

Research Mode

For structured multi-phase research:

  1. Switch to Research - Click the "Research" toggle button
  2. Select Manifest - Choose a research protocol from the dropdown
  3. Enter Subject - Type your research topic (company name, technology, etc.)
  4. Run Workflow - Click "Run" to start the multi-phase research
  5. Monitor Progress - Watch each phase complete with streaming output
  6. Follow Up - Ask questions about the research after completion

Session Management

Viewing Sessions

  • Session History - All sessions listed in sidebar, organized by date
  • Session Details - Click any session to view all phase outputs
  • Phase Expansion - Click phase headers to expand/collapse outputs

Session Resume

  • Incomplete Sessions - Sessions that didn't complete show "in_progress" status
  • Resume Button - Click "Resume" to continue from the last completed phase
  • Context Preservation - All previous phase outputs are used as context

Prompt Editing & Relaunch

  • View Prompts - Click "View Prompts" on any phase to see system/user prompts
  • Edit Mode - Click "Edit" to modify the prompts
  • Relaunch Phase - Click "Relaunch" to re-run with original or modified prompts
  • Reproducibility - Original prompts always preserved for reference

Project Organization

Group related research sessions:

  1. Create Project - Click "New Project" in the sidebar
  2. Name Project - Give it a descriptive name
  3. Add Sessions - Use session menu (⋯) → "Add to Project"
  4. View Project Sessions - Click project to expand and see all sessions
  5. Archive Projects - Archive completed projects to declutter

Export Options

  • Copy - Click copy icon to copy all output to clipboard
  • Export Markdown - Use menu to save as .md file
  • Print - Print-optimized view for physical copies
  • Per-Phase Export - Copy individual phase outputs

Keyboard Shortcuts

Action Shortcut
Send Message Enter
New Line in Message Shift+Enter
New Research Ctrl+N
Settings Ctrl+,
Export Ctrl+E

Troubleshooting

"API Key Invalid" Error

App Won't Start

Windows:

  • Ensure Windows 10/11 64-bit
  • Try running as Administrator
  • Check if WebView2 Runtime is installed

macOS:

  • Right-click the app → Open (to bypass Gatekeeper)
  • Check System Preferences → Security & Privacy if blocked
  • Ensure macOS 10.15 or later

Linux:

  • Check dependencies: libgtk-3-0, libwebkit2gtk-4.1-0
  • For AppImage: ensure FUSE is installed (sudo apt install fuse)
  • Run from terminal to see error messages

Research Hangs or Fails

  • Check your internet connection
  • Verify API key is valid and has credits
  • Check your provider's API status page
  • Try a different model or provider

Session Not Saving

  • Ensure you're logged in
  • Check disk space
  • Restart the application

Security

Known Advisories

Advisory Severity Platforms Status
GHSA-wrw7-89jp-8q8g Medium Linux only Upstream dependency

GHSA-wrw7-89jp-8q8g: A memory safety issue in the glib crate (v0.18.5) affects Linux builds. This is a transitive dependency from Tauri's GTK/WebKit stack and cannot be patched until the upstream gtk-rs ecosystem releases glib 0.20.0 compatibility. Windows and macOS builds are unaffected. The vulnerability requires malformed GVariant data from an untrusted source to exploit, which is not a vector in this application's design.

Reporting Security Issues

To report a security vulnerability, please open a private security advisory on GitHub or contact the maintainer directly.

Support

For technical support, contact your administrator.

Built with Tauri and React

About

FullIntel Agent: Sophisticated AI agent system with multi-LLM support, rate limiting, and circuit breaker patterns. Complete 60-test suite (Rust/Tauri).

Resources

Readme
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases 2

v0.1.1 Latest
Dec 2, 2025
+ 1 release

Packages

 
 
 

Contributors

Languages