Libraries¶

This document describes the third-party libraries and external tools used in Kartoza Screencaster and the rationale for their selection.

Core Libraries¶

Bubble Tea¶

github.com/charmbracelet/bubbletea

The Elm Architecture for terminal applications.

Version: v1.3.10

Purpose: Foundation for the entire TUI

Why chosen:

Elm Architecture provides predictable state management
Excellent performance and low memory usage
Active development and community
Works well with other Charm libraries

Usage:

type Model interface {
    Init() tea.Cmd
    Update(tea.Msg) (tea.Model, tea.Cmd)
    View() string
}

Lip Gloss¶

github.com/charmbracelet/lipgloss

Style definitions for terminal UIs.

Version: v1.1.1

Purpose: Consistent styling and layout

Why chosen:

Declarative styling API
Flexbox-like layout
Color support with degradation
Integrates seamlessly with Bubble Tea

Usage:

style := lipgloss.NewStyle().
    Bold(true).
    Foreground(lipgloss.Color("#FF9500")).
    Padding(1, 2)

rendered := style.Render("Hello, World!")

Bubbles¶

github.com/charmbracelet/bubbles

Common UI components for Bubble Tea.

Version: v0.21.1

Purpose: Reusable UI widgets

Components used:

Component	Usage
`textinput`	Title, description inputs
`textarea`	Multi-line description
`spinner`	Loading indicators
`progress`	Upload/processing bars
`key`	Keyboard handling
`table`	Recording history
`viewport`	Scrollable content areas

System Tray¶

Fyne Systray¶

fyne.io/systray

Cross-platform system tray library.

Version: v1.12.0

Purpose: System tray icon and menu for quick recording access

Why chosen:

Cross-platform (Linux, macOS, Windows)
Simple API
No CGO required on most platforms
Active maintenance

Image Processing¶

Go Terminal Image¶

github.com/blacktop/go-termimg

Display images in terminal emulators.

Version: v0.1.24

Purpose: Display webcam preview and thumbnails in the TUI

Why chosen:

Supports multiple terminal image protocols (Sixel, iTerm2, Kitty)
Automatic protocol detection
Fallback to ASCII art

Resize¶

github.com/nfnt/resize

Pure Go image resizing.

Version: v0.0.0-20180221191011-83c6a9932646

Purpose: Resize images for thumbnails and overlays

Why chosen:

Pure Go (no CGO)
Multiple resize algorithms (Lanczos, Bicubic, etc.)
Simple API

Text Processing¶

Fuzzy¶

github.com/sajari/fuzzy

Fuzzy string matching and spell checking.

Version: v1.0.0

Purpose: Spell checking for video titles and descriptions

Why chosen:

Fast fuzzy matching
Customizable word lists
Pure Go implementation

YouTube Integration¶

Google APIs¶

google.golang.org/api

Official Google API client libraries.

Version: v0.260.0

Packages used:

youtube/v3 - YouTube Data API
oauth2 - Authentication

Purpose: YouTube upload and playlist management

Usage:

import "google.golang.org/api/youtube/v3"

service, err := youtube.NewService(ctx, option.WithHTTPClient(client))
call := service.Videos.Insert([]string{"snippet", "status"}, video)
response, err := call.Media(file).Do()

OAuth2¶

golang.org/x/oauth2

OAuth 2.0 client implementation.

Version: v0.34.0

Purpose: Google authentication

Features used:

Authorization code flow
Token refresh
Secure token storage

CLI Framework¶

Cobra¶

github.com/spf13/cobra

CLI application framework.

Version: v1.10.2

Purpose: Command-line argument parsing

Why chosen:

Industry standard for Go CLIs
Automatic help generation
Subcommand support
Shell completion

Usage:

var rootCmd = &cobra.Command{
    Use:   "kvp",
    Short: "Kartoza Screencaster",
    Run:   runApp,
}

External Tools¶

These external command-line tools are used by Kartoza Screencaster for media processing:

FFmpeg¶

ffmpeg.org

Complete, cross-platform solution for recording, converting, and streaming audio and video.

Purpose: Video/audio encoding, merging, and filter processing

Features used:

Screen recording (x11grab, avfoundation, gdigrab)
Video encoding (H.264/libx264)
Audio encoding (AAC, PCM)
Video filters (scale, crop, overlay, drawtext)
Audio filters (loudnorm, pan)
Concatenation of video parts

License: LGPL/GPL

PipeWire (Linux)¶

pipewire.org

Low-latency audio/video server for Linux.

Purpose: Audio recording on Linux

Tools used:

pw-record - Audio capture
pw-cli - Device enumeration

Why chosen:

Modern replacement for PulseAudio
Lower latency
Better integration with Wayland

License: MIT

wf-recorder (Linux/Wayland)¶

github.com/ammen99/wf-recorder

Screen recorder for wlroots-based Wayland compositors.

Purpose: Screen recording on Wayland

Why chosen:

Native Wayland support
Hardware encoding support
Compositor-agnostic

License: MIT

Jivetalking¶

github.com/linuxmatters/jivetalking

Professional audio processing tool for podcast and screencast production.

Purpose: Advanced audio processing (when installed)

Features used:

Multi-pass audio analysis
Adaptive noise removal (anlmdn + compand)
LA-2A style optical compression
DS201-inspired gating
De-essing
EBU R128 loudness normalization

Processing pipeline:

Pass 1: Analysis (loudness, noise floor, speech characteristics)
Pass 2: Processing (high-pass, noise removal, gate, compression, de-esser)
Pass 3: Loudnorm measurement
Pass 4: Loudnorm application + click repair

Fallback: When jivetalking is not installed, Kartoza Screencaster falls back to basic FFmpeg loudnorm processing.

License: GPL-3.0

Dependency Graph¶

graph TD
    App[Kartoza Screencaster]

    subgraph "UI Layer"
        BT[bubbletea]
        LG[lipgloss]
        BB[bubbles]
    end

    subgraph "CLI"
        Cobra[cobra]
    end

    subgraph "YouTube"
        GAPI[google-api-go]
        OAuth[oauth2]
    end

    subgraph "System"
        Systray[fyne/systray]
    end

    subgraph "Image"
        TermImg[go-termimg]
        Resize[nfnt/resize]
    end

    subgraph "Text"
        Fuzzy[sajari/fuzzy]
    end

    subgraph "External Tools"
        FFmpeg[FFmpeg]
        PipeWire[PipeWire]
        WfRecorder[wf-recorder]
        Jivetalking[Jivetalking]
    end

    App --> BT
    App --> Cobra
    App --> Systray
    App --> TermImg
    App --> Fuzzy
    BT --> LG
    BT --> BB
    App --> GAPI
    GAPI --> OAuth
    TermImg --> Resize

    App -.->|shells out| FFmpeg
    App -.->|shells out| PipeWire
    App -.->|shells out| WfRecorder
    App -.->|shells out| Jivetalking

Version Management¶

go.mod¶

All Go dependencies are managed through Go modules:

module github.com/kartoza/kartoza-screencaster

go 1.24

require (
    fyne.io/systray v1.12.0
    github.com/blacktop/go-termimg v0.1.24
    github.com/charmbracelet/bubbles v0.21.1
    github.com/charmbracelet/bubbletea v1.3.10
    github.com/charmbracelet/lipgloss v1.1.1
    github.com/nfnt/resize v0.0.0-20180221191011-83c6a9932646
    github.com/sajari/fuzzy v1.0.0
    github.com/spf13/cobra v1.10.2
    golang.org/x/oauth2 v0.34.0
    google.golang.org/api v0.260.0
)

Updating Dependencies¶

# Update all
go get -u ./...

# Update specific library
go get -u github.com/charmbracelet/bubbletea

# Check for updates
go list -m -u all

# Tidy dependencies
go mod tidy

Selection Criteria¶

When choosing libraries, we consider:

Maintenance - Active development, responsive maintainers
Community - Usage, documentation, examples
Performance - Memory usage, CPU efficiency
API stability - Semantic versioning, deprecation policy
License - MIT, Apache 2.0, or similar permissive license
Size - Minimal dependencies, reasonable binary size

Alternatives Considered¶

TUI Framework¶

Library	Consideration	Decision
Bubble Tea	Elm architecture, excellent ecosystem	Selected
tview	More traditional, harder to test	Not selected
termui	Good for dashboards, less flexible	Not selected
tcell	Lower level, more boilerplate	Not selected

CLI Framework¶

Library	Consideration	Decision
Cobra	Industry standard, full featured	Selected
urfave/cli	Good but less ecosystem	Not selected
kong	Nice API but newer	Not selected

Audio Processing¶

Tool	Consideration	Decision
Jivetalking	Professional multi-pass processing, adaptive	Primary (if installed)
FFmpeg loudnorm	Basic two-pass normalization	Fallback
SoX	Good but less flexible filters	Not selected

License Summary¶

Go Libraries¶

Library	License
bubbletea	MIT
lipgloss	MIT
bubbles	MIT
fyne/systray	BSD-3-Clause
go-termimg	MIT
nfnt/resize	ISC
sajari/fuzzy	MIT
cobra	Apache 2.0
google-api-go	BSD-3-Clause
oauth2	BSD-3-Clause

External Tools¶

Tool	License
FFmpeg	LGPL/GPL
PipeWire	MIT
wf-recorder	MIT
Jivetalking	GPL-3.0

All Go dependencies use permissive open-source licenses compatible with the project's MIT license. External tools are called via subprocess and their licenses apply independently.