#!/usr/bin/osascript
--------------------------------------------------------------------------------
-- peepit.scpt - v1.0.0 "PeepIt Pro! 👀 → 📸 → 💾"
-- Enhanced screenshot capture with multi-window support and app discovery
-- PeepIt—screenshot got you! Now you see it, now it's saved.
--
-- IMPORTANT: This script uses non-interactive screencapture methods
-- Do NOT use flags like -o -W which require user interaction
-- Instead use -l<windowID> for specific window capture
--------------------------------------------------------------------------------
--#region Configuration Properties
property scriptInfoPrefix : "PeepIt 👀: "
property defaultScreenshotFormat : "png"
property captureDelay : 0.3
property windowActivationDelay : 0.2
property enhancedErrorReporting : true
property verboseLogging : false
property maxWindowTitleLength : 50
-- AI Analysis Configuration
property defaultVisionModel : "qwen2.5vl:7b"
-- Prioritized list of vision models (best to fallback)
property visionModelPriority : {"qwen2.5vl:7b", "llava:7b", "llava-phi3:3.8b", "minicpm-v:8b", "gemma3:4b", "llava:latest", "qwen2.5vl:3b", "llava:13b", "llava-llama3:8b"}
-- AI Provider Configuration
property aiProvider : "auto" -- "auto", "ollama", "claude"
property claudeModel : "sonnet" -- default Claude model alias
-- AI Analysis Timeout (90 seconds)
property aiAnalysisTimeout : 90
--#endregion Configuration Properties
--#region Helper Functions
on isValidPath(thePath)
if thePath is not "" and (thePath starts with "/") then
return true
end if
return false
end isValidPath
on getFileExtension(filePath)
set oldDelims to AppleScript's text item delimiters
set AppleScript's text item delimiters to "."
set pathParts to text items of filePath
set AppleScript's text item delimiters to oldDelims
if (count pathParts) > 1 then
return item -1 of pathParts
else
return ""
end if
end getFileExtension
on ensureDirectoryExists(dirPath)
try
do shell script "mkdir -p " & quoted form of dirPath
return true
on error
return false
end try
end ensureDirectoryExists
on sanitizeFilename(filename)
-- Replace problematic characters for filenames
set filename to my replaceText(filename, "/", "_")
set filename to my replaceText(filename, ":", "_")
set filename to my replaceText(filename, "*", "_")
set filename to my replaceText(filename, "?", "_")
set filename to my replaceText(filename, "\"", "_")
set filename to my replaceText(filename, "<", "_")
set filename to my replaceText(filename, ">", "_")
set filename to my replaceText(filename, "|", "_")
if (length of filename) > maxWindowTitleLength then
set filename to text 1 thru maxWindowTitleLength of filename
end if
return filename
end sanitizeFilename
on sanitizeAppName(appName)
-- Create model-friendly app names (lowercase, underscores, no spaces)
set appName to my replaceText(appName, " ", "_")
set appName to my replaceText(appName, ".", "_")
set appName to my replaceText(appName, "-", "_")
set appName to my replaceText(appName, "/", "_")
set appName to my replaceText(appName, ":", "_")
set appName to my replaceText(appName, "*", "_")
set appName to my replaceText(appName, "?", "_")
set appName to my replaceText(appName, "\"", "_")
set appName to my replaceText(appName, "<", "_")
set appName to my replaceText(appName, ">", "_")
set appName to my replaceText(appName, "|", "_")
-- Convert to lowercase using shell
try
set appName to do shell script "echo " & quoted form of appName & " | tr '[:upper:]' '[:lower:]'"
on error
-- Fallback if shell command fails
end try
-- Limit length for readability
if (length of appName) > 20 then
set appName to text 1 thru 20 of appName
end if
return appName
end sanitizeAppName
on replaceText(theText, searchStr, replaceStr)
set oldDelims to AppleScript's text item delimiters
set AppleScript's text item delimiters to searchStr
set textItems to text items of theText
set AppleScript's text item delimiters to replaceStr
set newText to textItems as text
set AppleScript's text item delimiters to oldDelims
return newText
end replaceText
on formatErrorMessage(errorType, errorMsg, context)
if enhancedErrorReporting then
set formattedMsg to scriptInfoPrefix & errorType & ": " & errorMsg
if context is not "" then
set formattedMsg to formattedMsg & " (Context: " & context & ")"
end if
return formattedMsg
else
return scriptInfoPrefix & errorMsg
end if
end formatErrorMessage
on logVerbose(message)
if verboseLogging then
log "🔍 " & message
end if
end logVerbose
on trimWhitespace(theText)
set whitespaceChars to {" ", tab}
set newText to theText
repeat while (newText is not "") and (character 1 of newText is in whitespaceChars)
if (length of newText) > 1 then
set newText to text 2 thru -1 of newText
else
set newText to ""
end if
end repeat
repeat while (newText is not "") and (character -1 of newText is in whitespaceChars)
if (length of newText) > 1 then
set newText to text 1 thru -2 of newText
else
set newText to ""
end if
end repeat
return newText
end trimWhitespace
on formatCaptureOutput(outputPath, appName, mode, isQuiet)
if isQuiet then
return outputPath
else
set msg to scriptInfoPrefix & "Screenshot captured successfully! 📸" & linefeed
set msg to msg & "• File: " & outputPath & linefeed
set msg to msg & "• App: " & appName & linefeed
set msg to msg & "• Mode: " & mode
return msg
end if
end formatCaptureOutput
on formatMultiOutput(capturedFiles, appName, isQuiet)
if isQuiet then
-- Just return paths separated by newlines
set paths to ""
repeat with fileInfo in capturedFiles
set filePath to item 1 of fileInfo
set paths to paths & filePath & linefeed
end repeat
return paths
else
set windowCount to count of capturedFiles
set msg to scriptInfoPrefix & "Multi-window capture successful! Captured " & windowCount & " window(s) for " & appName & ":" & linefeed
repeat with fileInfo in capturedFiles
set filePath to item 1 of fileInfo
set winTitle to item 2 of fileInfo
set msg to msg & " 📸 " & filePath & " → \"" & winTitle & "\"" & linefeed
end repeat
return msg
end if
end formatMultiOutput
on formatMultiWindowAnalysis(capturedFiles, analysisResults, appName, question, model, isQuiet)
if isQuiet then
-- In quiet mode, return condensed results
set output to ""
repeat with result in analysisResults
set winTitle to windowTitle of result
set answer to answer of result
set output to output & scriptInfoPrefix & "Window \"" & winTitle & "\": " & answer & linefeed
end repeat
return output
else
-- Full formatted output
set windowCount to count of capturedFiles
set msg to scriptInfoPrefix & "Multi-window AI Analysis Complete! 🤖" & linefeed & linefeed
set msg to msg & "📸 App: " & appName & " (" & windowCount & " windows)" & linefeed
set msg to msg & "❓ Question: " & question & linefeed
set msg to msg & "🤖 Model: " & model & linefeed & linefeed
set msg to msg & "💬 Results for each window:" & linefeed & linefeed
set windowNum to 1
repeat with result in analysisResults
set winTitle to windowTitle of result
set winIndex to windowIndex of result
set answer to answer of result
set success to success of result
set msg to msg & "🪟 Window " & windowNum & ": \"" & winTitle & "\"" & linefeed
if success then
set msg to msg & answer & linefeed & linefeed
else
set msg to msg & "⚠️ Analysis failed: " & answer & linefeed & linefeed
end if
set windowNum to windowNum + 1
end repeat
-- Add timing info if available
set msg to msg & scriptInfoPrefix & "Analysis of " & windowCount & " windows complete."
return msg
end if
end formatMultiWindowAnalysis
--#endregion Helper Functions
--#region AI Analysis Functions
on checkOllamaAvailable()
try
-- Check if ollama command exists
do shell script "ollama --version >/dev/null 2>&1"
-- Check if ollama service is running by testing API
do shell script "curl -s http://localhost:11434/api/tags >/dev/null 2>&1"
return true
on error
return false
end try
end checkOllamaAvailable
on checkClaudeAvailable()
try
-- Check if claude command exists
do shell script "claude --version >/dev/null 2>&1"
return true
on error
return false
end try
end checkClaudeAvailable
on getAvailableVisionModels()
set availableModels to {}
try
set ollamaList to do shell script "ollama list 2>/dev/null | tail -n +2 | awk '{print $1}' | grep -v '^$'"
set modelLines to paragraphs of ollamaList
repeat with modelLine in modelLines
set modelName to contents of modelLine
if modelName is not "" then
set end of availableModels to modelName
end if
end repeat
on error
-- Return empty list if ollama list fails
end try
return availableModels
end getAvailableVisionModels
on findBestVisionModel(requestedModel)
my logVerbose("Finding best vision model, requested: " & requestedModel)
set availableModels to my getAvailableVisionModels()
my logVerbose("Available models: " & (availableModels as string))
-- If specific model requested and available, use it
if requestedModel is not defaultVisionModel then
repeat with availModel in availableModels
if contents of availModel is requestedModel then
my logVerbose("Using requested model: " & requestedModel)
return requestedModel
end if
end repeat
-- Requested model not found, will fall back to priority list
my logVerbose("Requested model '" & requestedModel & "' not found, checking priority list")
end if
-- Find best available model from priority list
repeat with priorityModel in visionModelPriority
repeat with availModel in availableModels
if contents of availModel is contents of priorityModel then
my logVerbose("Using priority model: " & contents of priorityModel)
return contents of priorityModel
end if
end repeat
end repeat
-- No priority models available, use first available vision model
repeat with availModel in availableModels
set modelName to contents of availModel
if modelName contains "llava" or modelName contains "qwen" or modelName contains "gemma" or modelName contains "minicpm" then
my logVerbose("Using first available vision model: " & modelName)
return modelName
end if
end repeat
-- No vision models found
return ""
end findBestVisionModel
on getOllamaInstallInstructions()
set instructions to scriptInfoPrefix & "AI Analysis requires Ollama with a vision model." & linefeed & linefeed
set instructions to instructions & "🚀 Quick Setup:" & linefeed
set instructions to instructions & "1. Install Ollama: curl -fsSL https://ollama.ai/install.sh | sh" & linefeed
set instructions to instructions & "2. Pull a vision model: ollama pull " & defaultVisionModel & linefeed
set instructions to instructions & "3. Models are ready to use!" & linefeed & linefeed
set instructions to instructions & "💡 Recommended models:" & linefeed
set instructions to instructions & " • qwen2.5vl:7b (6GB) - Best doc/chart understanding" & linefeed
set instructions to instructions & " • llava:7b (4.7GB) - Solid all-rounder" & linefeed
set instructions to instructions & " • llava-phi3:3.8b (2.9GB) - Tiny but chatty" & linefeed
set instructions to instructions & " • minicpm-v:8b (5.5GB) - Killer OCR" & linefeed & linefeed
set instructions to instructions & "Then retry your PeepIt command with --ask or --analyze!"
return instructions
end getOllamaInstallInstructions
on analyzeImageWithOllama(imagePath, question, requestedModel)
my logVerbose("Analyzing image with AI: " & imagePath)
my logVerbose("Requested model: " & requestedModel)
my logVerbose("Question: " & question)
-- Record start time
set startTime to do shell script "date +%s"
-- Check if Ollama is available
if not my checkOllamaAvailable() then
return my formatErrorMessage("Ollama Error", "Ollama is not installed or not in PATH." & linefeed & linefeed & my getOllamaInstallInstructions(), "ollama unavailable")
end if
-- Find best available vision model
set modelToUse to my findBestVisionModel(requestedModel)
if modelToUse is "" then
return my formatErrorMessage("Model Error", "No vision models found." & linefeed & linefeed & my getOllamaInstallInstructions(), "no vision models")
end if
-- Use ollama run command with proper vision model syntax
try
my logVerbose("Using model: " & modelToUse)
-- For vision models, we need to use the API approach or different command structure
-- Let's use a simpler approach with base64 and API
-- First check if image is too large and compress if needed
set imageSize to do shell script "wc -c < " & quoted form of imagePath
set imageSizeBytes to imageSize as number
my logVerbose("Image size: " & imageSize & " bytes")
set processedImagePath to imagePath
if imageSizeBytes > 5000000 then -- 5MB threshold
my logVerbose("Image is large (" & imageSize & " bytes), creating compressed version for AI")
set compressedPath to "/tmp/peepit_ai_compressed.png"
-- Use sips to resize and compress image for AI analysis
do shell script "sips -Z 2048 -s format png " & quoted form of imagePath & " --out " & quoted form of compressedPath
set processedImagePath to compressedPath
end if
set base64Image to do shell script "base64 -i " & quoted form of processedImagePath & " | tr -d '\\n'"
set jsonPayload to "{\"model\":\"" & modelToUse & "\",\"prompt\":\"" & my escapeJSON(question) & "\",\"images\":[\"" & base64Image & "\"],\"stream\":false}"
my logVerbose("Running API call to Ollama")
my logVerbose("JSON payload size: " & (length of jsonPayload) & " characters")
my logVerbose("Base64 image size: " & (length of base64Image) & " characters")
-- Write JSON to temporary file using AppleScript file writing to avoid shell limitations
set jsonTempFile to "/tmp/peepit_ollama_request.json"
try
set fileRef to open for access (POSIX file jsonTempFile) with write permission
set eof of fileRef to 0
write jsonPayload to fileRef
close access fileRef
on error
try
close access fileRef
end try
end try
-- Add timeout to curl command (60 seconds)
set curlCmd to "curl -s -X POST http://localhost:11434/api/generate -H 'Content-Type: application/json' -d @" & quoted form of jsonTempFile & " --max-time " & aiAnalysisTimeout
set response to do shell script curlCmd
-- Parse JSON response
set responseStart to (offset of "\"response\":\"" in response) + 12
if responseStart > 12 then
set responseEnd to responseStart
set inEscape to false
repeat with i from responseStart to (length of response)
set char to character i of response
if inEscape then
set inEscape to false
else if char is "\\" then
set inEscape to true
else if char is "\"" then
set responseEnd to i - 1
exit repeat
end if
end repeat
set aiResponse to text responseStart thru responseEnd of response
-- Unescape JSON
set aiResponse to my replaceText(aiResponse, "\\n", linefeed)
set aiResponse to my replaceText(aiResponse, "\\\"", "\"")
set aiResponse to my replaceText(aiResponse, "\\\\", "\\")
else
error "Could not parse response: " & response
end if
-- Calculate elapsed time
set endTime to do shell script "date +%s"
set elapsedTime to (endTime as number) - (startTime as number)
-- Simple formatting - just show seconds
set elapsedTimeFormatted to elapsedTime as string
set resultMsg to scriptInfoPrefix & "AI Analysis Complete! 🤖" & linefeed & linefeed
set resultMsg to resultMsg & "📸 Image: " & imagePath & linefeed
set resultMsg to resultMsg & "❓ Question: " & question & linefeed
set resultMsg to resultMsg & "🤖 Model: " & modelToUse & linefeed & linefeed
set resultMsg to resultMsg & "💬 Answer:" & linefeed & aiResponse & linefeed & linefeed
set resultMsg to resultMsg & scriptInfoPrefix & "Analysis via " & modelToUse & " took " & elapsedTimeFormatted & " sec."
return resultMsg
on error errMsg
-- Calculate elapsed time even on error
set endTime to do shell script "date +%s"
set elapsedTime to (endTime as number) - (startTime as number)
if errMsg contains "curl" and (errMsg contains "timed out" or errMsg contains "timeout" or elapsedTime ≥ aiAnalysisTimeout) then
return my formatErrorMessage("Timeout Error", "AI analysis timed out after " & aiAnalysisTimeout & " seconds." & linefeed & linefeed & "The model '" & modelToUse & "' may be too large or slow for your system." & linefeed & linefeed & "Try:" & linefeed & "• Using a smaller model (e.g., llava-phi3:3.8b)" & linefeed & "• Checking if Ollama is responding: ollama list" & linefeed & "• Restarting Ollama service", "timeout")
else if errMsg contains "model" and errMsg contains "not found" then
return my formatErrorMessage("Model Error", "Model '" & modelToUse & "' not found." & linefeed & linefeed & "Install it with: ollama pull " & modelToUse & linefeed & linefeed & my getOllamaInstallInstructions(), "model not found")
else
return my formatErrorMessage("Analysis Error", "Failed to analyze image: " & errMsg & linefeed & linefeed & "Make sure Ollama is running and the model is available.", "ollama execution")
end if
end try
end analyzeImageWithAI
on escapeJSON(inputText)
set escapedText to my replaceText(inputText, "\\", "\\\\")
set escapedText to my replaceText(escapedText, "\"", "\\\"")
set escapedText to my replaceText(escapedText, linefeed, "\\n")
set escapedText to my replaceText(escapedText, return, "\\n")
set escapedText to my replaceText(escapedText, tab, "\\t")
return escapedText
end escapeJSON
on analyzeImageWithClaude(imagePath, question, modelAlias)
my logVerbose("Analyzing image with Claude: " & imagePath)
my logVerbose("Model: " & modelAlias)
my logVerbose("Question: " & question)
-- Record start time
set startTime to do shell script "date +%s"
-- Check if Claude is available
if not my checkClaudeAvailable() then
return my formatErrorMessage("Claude Error", "Claude CLI is not installed." & linefeed & linefeed & "Install it from: https://claude.ai/code", "claude unavailable")
end if
-- Get Claude version
set claudeVersion to ""
try
set claudeVersion to do shell script "claude --version 2>/dev/null | head -1"
on error
set claudeVersion to "unknown"
end try
try
-- Note: Claude CLI doesn't support direct image file analysis
-- This is a limitation of the current Claude CLI implementation
set errorMsg to "Claude CLI currently doesn't support direct image file analysis." & linefeed & linefeed
set errorMsg to errorMsg & "Claude can analyze images through:" & linefeed
set errorMsg to errorMsg & "• Copy/paste images in interactive mode" & linefeed
set errorMsg to errorMsg & "• MCP (Model Context Protocol) integrations" & linefeed & linefeed
set errorMsg to errorMsg & "For automated image analysis, please use Ollama with vision models instead."
-- Calculate elapsed time even for error
set endTime to do shell script "date +%s"
set elapsedTime to (endTime as number) - (startTime as number)
set elapsedTimeFormatted to elapsedTime as string
set errorMsg to errorMsg & linefeed & linefeed & scriptInfoPrefix & "Claude " & claudeVersion & " check took " & elapsedTimeFormatted & " sec."
return my formatErrorMessage("Claude Limitation", errorMsg, "feature not supported")
on error errMsg
return my formatErrorMessage("Claude Analysis Error", "Failed to analyze image with Claude: " & errMsg, "claude execution")
end try
end analyzeImageWithClaude
on analyzeImageWithAI(imagePath, question, requestedModel, requestedProvider)
my logVerbose("Starting AI analysis with smart provider selection")
my logVerbose("Requested provider: " & requestedProvider)
-- Determine which AI provider to use
set ollamaAvailable to my checkOllamaAvailable()
set claudeAvailable to my checkClaudeAvailable()
my logVerbose("Ollama available: " & ollamaAvailable)
my logVerbose("Claude available: " & claudeAvailable)
-- If neither is available, provide helpful error
if not ollamaAvailable and not claudeAvailable then
set errorMsg to "Neither Ollama nor Claude CLI is installed." & linefeed & linefeed
set errorMsg to errorMsg & "Install one of these AI providers:" & linefeed & linefeed
set errorMsg to errorMsg & "🤖 Ollama (local, privacy-focused):" & linefeed
set errorMsg to errorMsg & my getOllamaInstallInstructions() & linefeed & linefeed
set errorMsg to errorMsg & "☁️ Claude CLI (cloud-based):" & linefeed
set errorMsg to errorMsg & "Install from: https://claude.ai/code"
return my formatErrorMessage("No AI Provider", errorMsg, "no ai provider")
end if
-- Smart selection based on availability and preference
if requestedProvider is "ollama" and ollamaAvailable then
return my analyzeImageWithOllama(imagePath, question, requestedModel)
else if requestedProvider is "claude" and claudeAvailable then
return my analyzeImageWithClaude(imagePath, question, requestedModel)
else if requestedProvider is "auto" then
-- Auto mode: prefer Ollama, fallback to Claude
if ollamaAvailable then
return my analyzeImageWithOllama(imagePath, question, requestedModel)
else if claudeAvailable then
return my analyzeImageWithClaude(imagePath, question, requestedModel)
end if
else
-- Requested provider not available, try the other one
if ollamaAvailable then
my logVerbose("Requested provider not available, using Ollama instead")
return my analyzeImageWithOllama(imagePath, question, requestedModel)
else if claudeAvailable then
my logVerbose("Requested provider not available, using Claude instead")
return my analyzeImageWithClaude(imagePath, question, requestedModel)
end if
end if
-- Should never reach here
return my formatErrorMessage("Provider Error", "Unable to determine AI provider", "provider selection")
end analyzeImageWithAI
--#endregion AI Analysis Functions
--#region App Discovery Functions
on listRunningApps()
set appList to {}
try
tell application "System Events"
repeat with proc in (every application process whose background only is false)
try
set appName to name of proc
set bundleID to bundle identifier of proc
set windowCount to count of windows of proc
set windowTitles to {}
if windowCount > 0 then
repeat with win in windows of proc
try
set winTitle to title of win
if winTitle is not "" then
set end of windowTitles to winTitle
end if
on error
-- Skip windows without accessible titles
end try
end repeat
end if
set end of appList to {appName:appName, bundleID:bundleID, windowCount:windowCount, windowTitles:windowTitles}
on error
-- Skip apps we can't access
end try
end repeat
end tell
on error errMsg
return my formatErrorMessage("Discovery Error", "Failed to enumerate running applications: " & errMsg, "app enumeration")
end try
return appList
end listRunningApps
on formatAppList(appList)
if appList starts with scriptInfoPrefix then
return appList -- Error message
end if
set output to scriptInfoPrefix & "Running Applications:" & linefeed & linefeed
repeat with appInfo in appList
set appName to appName of appInfo
set bundleID to bundleID of appInfo
set windowCount to windowCount of appInfo
set windowTitles to windowTitles of appInfo
set output to output & "• " & appName & " (" & bundleID & ")" & linefeed
set output to output & " Windows: " & windowCount
if windowCount > 0 and (count of windowTitles) > 0 then
set output to output & linefeed
repeat with winTitle in windowTitles
set output to output & " - \"" & winTitle & "\"" & linefeed
end repeat
else
set output to output & linefeed
end if
set output to output & linefeed
end repeat
return output
end formatAppList
--#endregion App Discovery Functions
--#region App Resolution Functions
on resolveAppIdentifier(appIdentifier)
my logVerbose("Resolving app identifier: " & appIdentifier)
-- First try as bundle ID
try
tell application "System Events"
set bundleApps to (every application process whose bundle identifier is appIdentifier)
if (count bundleApps) > 0 then
set targetApp to item 1 of bundleApps
set appName to name of targetApp
my logVerbose("Found running app by bundle ID: " & appName)
return {appName:appName, bundleID:appIdentifier, isRunning:true, resolvedBy:"bundle_id"}
end if
end tell
on error
my logVerbose("Bundle ID lookup failed, trying as app name")
end try
-- Try as application name for running apps
try
tell application "System Events"
set nameApps to (every application process whose name is appIdentifier)
if (count nameApps) > 0 then
set targetApp to item 1 of nameApps
try
set bundleID to bundle identifier of targetApp
on error
set bundleID to ""
end try
my logVerbose("Found running app by name: " & appIdentifier)
return {appName:appIdentifier, bundleID:bundleID, isRunning:true, resolvedBy:"app_name"}
end if
end tell
on error
my logVerbose("App name lookup failed for running processes")
end try
-- Try to find the app in /Applications (not running)
try
set appPath to "/Applications/" & appIdentifier & ".app"
tell application "System Events"
if exists file appPath then
try
set bundleID to bundle identifier of file appPath
on error
set bundleID to ""
end try
my logVerbose("Found app in /Applications: " & appIdentifier)
return {appName:appIdentifier, bundleID:bundleID, isRunning:false, resolvedBy:"applications_folder"}
end if
end tell
on error
my logVerbose("/Applications lookup failed")
end try
-- If it looks like a bundle ID, try launching it directly
if appIdentifier contains "." then
try
tell application "System Events"
launch application id appIdentifier
delay windowActivationDelay
set bundleApps to (every application process whose bundle identifier is appIdentifier)
if (count bundleApps) > 0 then
set targetApp to item 1 of bundleApps
set appName to name of targetApp
my logVerbose("Successfully launched app by bundle ID: " & appName)
return {appName:appName, bundleID:appIdentifier, isRunning:true, resolvedBy:"bundle_id_launch"}
end if
end tell
on error errMsg
my logVerbose("Bundle ID launch failed: " & errMsg)
end try
end if
return missing value
end resolveAppIdentifier
on getAppWindows(appName)
set windowInfo to {}
set windowCount to 0
set accessibleWindows to 0
try
tell application "System Events"
tell process appName
set windowCount to count of windows
repeat with i from 1 to windowCount
try
set win to window i
set winTitle to title of win
if winTitle is "" then set winTitle to "Untitled Window " & i
set end of windowInfo to {winTitle, i}
set accessibleWindows to accessibleWindows + 1
on error
set end of windowInfo to {("Window " & i), i}
end try
end repeat
end tell
end tell
on error errMsg
my logVerbose("Failed to get windows for " & appName & ": " & errMsg)
return {windowInfo:windowInfo, totalWindows:0, accessibleWindows:0, errorMsg:errMsg}
end try
return {windowInfo:windowInfo, totalWindows:windowCount, accessibleWindows:accessibleWindows, errorMsg:""}
end getAppWindows
on getAppWindowStatus(appName)
set windowStatus to my getAppWindows(appName)
set windowInfo to windowInfo of windowStatus
set totalWindows to totalWindows of windowStatus
set accessibleWindows to accessibleWindows of windowStatus
set windowError to errorMsg of windowStatus
if windowError is not "" then
return my formatErrorMessage("Window Access Error", "Cannot access windows for app '" & appName & "': " & windowError & ". The app may not be running or may not have accessibility permissions.", "window enumeration")
end if
if totalWindows = 0 then
return my formatErrorMessage("No Windows Error", "App '" & appName & "' is running but has no windows open. PeepIt needs at least one window to capture. Please open a window in " & appName & " and try again.", "zero windows")
end if
if accessibleWindows = 0 and totalWindows > 0 then
return my formatErrorMessage("Window Access Error", "App '" & appName & "' has " & totalWindows & " window(s) but none are accessible. This may require accessibility permissions in System Preferences > Security & Privacy > Accessibility.", "accessibility required")
end if
-- Success case
set statusMsg to "App '" & appName & "' has " & totalWindows & " window(s)"
if accessibleWindows < totalWindows then
set statusMsg to statusMsg & " (" & accessibleWindows & " accessible)"
end if
return {status:"success", message:statusMsg, windowInfo:windowInfo, totalWindows:totalWindows, accessibleWindows:accessibleWindows}
end getAppWindowStatus
on bringAppToFront(appInfo)
set appName to appName of appInfo
set isRunning to isRunning of appInfo
my logVerbose("Bringing app to front: " & appName & " (running: " & isRunning & ")")
-- Skip app focus for fullscreen mode
if appName is "fullscreen" then
my logVerbose("Fullscreen mode - skipping app focus")
return ""
end if
if not isRunning then
try
tell application appName to activate
delay windowActivationDelay
on error errMsg
return my formatErrorMessage("Activation Error", "Failed to launch app '" & appName & "': " & errMsg, "app launch")
end try
else
try
tell application "System Events"
tell process appName
set frontmost to true
end tell
end tell
delay windowActivationDelay
on error errMsg
return my formatErrorMessage("Focus Error", "Failed to bring app '" & appName & "' to front: " & errMsg, "app focus")
end try
end if
return ""
end bringAppToFront
--#endregion App Resolution Functions
--#region Screenshot Functions
on captureScreenshot(outputPath, captureMode, appName)
my logVerbose("Capturing screenshot to: " & outputPath & " (mode: " & captureMode & ")")
-- Ensure output directory exists
set outputDir to do shell script "dirname " & quoted form of outputPath
if not my ensureDirectoryExists(outputDir) then
return my formatErrorMessage("Directory Error", "Could not create output directory: " & outputDir, "directory creation")
end if
-- Wait for capture delay
delay captureDelay
-- Determine screenshot format
set fileExt to my getFileExtension(outputPath)
if fileExt is "" then
set fileExt to defaultScreenshotFormat
set outputPath to outputPath & "." & fileExt
end if
-- Build screencapture command based on mode
set screencaptureCmd to "screencapture -x"
if captureMode is "window" and appName is not "fullscreen" then
-- IMPORTANT: Do NOT use -o -W flags as they require user interaction!
-- Instead, capture the window using its bounds (position and size)
try
tell application "System Events"
tell process appName
set winPosition to position of window 1
set winSize to size of window 1
end tell
end tell
set x to item 1 of winPosition
set y to item 2 of winPosition
set w to item 1 of winSize
set h to item 2 of winSize
-- Use -R flag to capture a specific rectangle
set screencaptureCmd to screencaptureCmd & " -R" & x & "," & y & "," & w & "," & h
my logVerbose("Capturing window bounds for " & appName & ": " & x & "," & y & "," & w & "," & h)
on error errMsg
-- Fallback to full screen if we can't get window bounds
my logVerbose("Could not get window bounds for " & appName & ", error: " & errMsg)
log scriptInfoPrefix & "Warning: Could not capture window bounds for " & appName & ", using full screen capture instead"
end try
end if
-- Add format flag if not PNG (default)
if fileExt is not "png" then
set screencaptureCmd to screencaptureCmd & " -t " & fileExt
end if
-- Add output path
set screencaptureCmd to screencaptureCmd & " " & quoted form of outputPath
-- Capture screenshot
try
my logVerbose("Running: " & screencaptureCmd)
do shell script screencaptureCmd
-- Verify file was created
try
do shell script "test -f " & quoted form of outputPath
return outputPath
on error
return my formatErrorMessage("Capture Error", "Screenshot file was not created at: " & outputPath, "file verification")
end try
on error errMsg number errNum
-- Enhanced error handling for common screencapture issues
if errMsg contains "not authorized" or errMsg contains "Screen Recording" then
return my formatErrorMessage("Permission Error", "Screen Recording permission required. Please go to System Preferences > Security & Privacy > Screen Recording and add your terminal app to the allowed list. Then restart your terminal and try again.", "screen recording permission")
else if errMsg contains "No such file" then
return my formatErrorMessage("Path Error", "Cannot create screenshot at '" & outputPath & "'. Check that the directory exists and you have write permissions.", "file creation")
else if errMsg contains "Permission denied" then
return my formatErrorMessage("Permission Error", "Permission denied writing to '" & outputPath & "'. Check file/directory permissions or try a different location like /tmp/", "write permission")
else
return my formatErrorMessage("Capture Error", "screencapture failed: " & errMsg & ". This may be due to permissions, disk space, or system restrictions.", "error " & errNum)
end if
end try
end captureScreenshot
on captureWindowByIndex(outputPath, appName, winIndex)
my logVerbose("Capturing window " & winIndex & " of " & appName & " to: " & outputPath)
-- Ensure output directory exists
set outputDir to do shell script "dirname " & quoted form of outputPath
if not my ensureDirectoryExists(outputDir) then
return my formatErrorMessage("Directory Error", "Could not create output directory: " & outputDir, "directory creation")
end if
-- Wait for capture delay
delay captureDelay
-- Determine screenshot format
set fileExt to my getFileExtension(outputPath)
if fileExt is "" then
set fileExt to defaultScreenshotFormat
set outputPath to outputPath & "." & fileExt
end if
-- Build screencapture command
set screencaptureCmd to "screencapture -x"
-- Get window bounds for specific window
try
tell application "System Events"
tell process appName
set winPosition to position of window winIndex
set winSize to size of window winIndex
end tell
end tell
set x to item 1 of winPosition
set y to item 2 of winPosition
set w to item 1 of winSize
set h to item 2 of winSize
-- Use -R flag to capture a specific rectangle
set screencaptureCmd to screencaptureCmd & " -R" & x & "," & y & "," & w & "," & h
my logVerbose("Capturing window " & winIndex & " bounds: " & x & "," & y & "," & w & "," & h)
on error errMsg
-- Fallback to full screen if we can't get window bounds
my logVerbose("Could not get bounds for window " & winIndex & ", error: " & errMsg)
return my formatErrorMessage("Window Error", "Could not get bounds for window " & winIndex & " of " & appName, "window bounds")
end try
-- Add format flag if not PNG (default)
if fileExt is not "png" then
set screencaptureCmd to screencaptureCmd & " -t " & fileExt
end if
-- Add output path
set screencaptureCmd to screencaptureCmd & " " & quoted form of outputPath
-- Capture screenshot
try
my logVerbose("Running: " & screencaptureCmd)
do shell script screencaptureCmd
-- Verify file was created
try
do shell script "test -f " & quoted form of outputPath
return outputPath
on error
return my formatErrorMessage("Capture Error", "Screenshot file was not created at: " & outputPath, "file verification")
end try
on error errMsg
return my formatErrorMessage("Capture Error", "Failed to capture window: " & errMsg, "screencapture")
end try
end captureWindowByIndex
on captureMultipleWindows(appName, baseOutputPath)
-- Get detailed window status first
set windowStatus to my getAppWindowStatus(appName)
-- Check if it's an error (string) or success (record)
try
set statusClass to class of windowStatus
if statusClass is text or statusClass is string then
-- It's an error message
return windowStatus
end if
on error
-- Assume it's a record and continue
end try
-- Extract window info from successful status
set windowInfo to windowInfo of windowStatus
set totalWindows to totalWindows of windowStatus
set accessibleWindows to accessibleWindows of windowStatus
set capturedFiles to {}
my logVerbose("Multi-window capture: " & totalWindows & " total, " & accessibleWindows & " accessible")
if (count of windowInfo) = 0 then
return my formatErrorMessage("Multi-Window Error", "App '" & appName & "' has no accessible windows for multi-window capture. Try using single screenshot mode instead, or ensure the app has open windows.", "no accessible windows")
end if
-- Get base path components
set outputDir to do shell script "dirname " & quoted form of baseOutputPath
set baseName to do shell script "basename " & quoted form of baseOutputPath
set fileExt to my getFileExtension(baseName)
if fileExt is not "" then
set baseNameNoExt to text 1 thru -((length of fileExt) + 2) of baseName
else
set baseNameNoExt to baseName
set fileExt to defaultScreenshotFormat
end if
my logVerbose("Capturing " & (count of windowInfo) & " windows for " & appName)
repeat with winInfo in windowInfo
set winTitle to item 1 of winInfo
set winIndex to item 2 of winInfo
set sanitizedTitle to my sanitizeAppName(winTitle)
set windowFileName to baseNameNoExt & "_window_" & winIndex & "_" & sanitizedTitle & "." & fileExt
set windowOutputPath to outputDir & "/" & windowFileName
-- Focus the specific window first
try
tell application "System Events"
tell process appName
set frontmost to true
tell window winIndex
perform action "AXRaise"
end tell
end tell
end tell
delay 0.1
on error
my logVerbose("Could not focus window " & winIndex & ", continuing anyway")
end try
-- Capture the specific window using a custom method
set captureResult to my captureWindowByIndex(windowOutputPath, appName, winIndex)
if captureResult starts with scriptInfoPrefix then
-- Error occurred, but continue with other windows
my logVerbose("Failed to capture window " & winIndex & ": " & captureResult)
else
set end of capturedFiles to {captureResult, winTitle, winIndex}
end if
end repeat
return capturedFiles
end captureMultipleWindows
--#endregion Screenshot Functions
--#region Main Script Logic (on run)
on run argv
set appSpecificErrorOccurred to false
try
my logVerbose("Starting PeepIt v2.0.0")
set argCount to count argv
-- Initialize all variables
set command to "" -- "capture", "analyze", "list", "help"
set appIdentifier to ""
set outputPath to ""
set outputSpecified to false
set captureMode to "" -- will be determined
set forceFullscreen to false
set multiWindow to false
set analyzeMode to false
set analysisQuestion to ""
set visionModel to defaultVisionModel
set requestedProvider to aiProvider
set outputFormat to ""
set quietMode to false
-- Handle no arguments - default to fullscreen
if argCount = 0 then
set command to "capture"
set forceFullscreen to true
else
-- Check first argument for commands
set firstArg to item 1 of argv
if firstArg is "list" or firstArg is "ls" then
return my formatAppList(my listRunningApps())
else if firstArg is "help" or firstArg is "-h" or firstArg is "--help" then
return my usageText()
else if firstArg is "analyze" then
set command to "analyze"
-- analyze command requires at least image and question
if argCount < 3 then
return my formatErrorMessage("Argument Error", "analyze command requires: analyze <image> \"question\"" & linefeed & linefeed & my usageText(), "validation")
end if
set appIdentifier to item 2 of argv -- actually the image path
set analysisQuestion to item 3 of argv
set analyzeMode to true
else
-- Regular capture command
set command to "capture"
-- Check if first arg is a flag or app name
if not (firstArg starts with "-") then
set appIdentifier to firstArg
end if
end if
end if
-- Parse remaining arguments
set i to 1
if command is "analyze" then set i to 4 -- Skip "analyze image question"
if command is "capture" and appIdentifier is not "" then set i to 2 -- Skip app name
repeat while i ≤ argCount
set arg to item i of argv
-- Handle flags with values
if arg is "--output" or arg is "-o" then
if i < argCount then
set i to i + 1
set outputPath to item i of argv
set outputSpecified to true
else
return my formatErrorMessage("Argument Error", arg & " requires a path parameter", "validation")
end if
else if arg is "--ask" or arg is "-a" then
if i < argCount then
set i to i + 1
set analysisQuestion to item i of argv
set analyzeMode to true
else
return my formatErrorMessage("Argument Error", arg & " requires a question parameter", "validation")
end if
else if arg is "--model" then
if i < argCount then
set i to i + 1
set visionModel to item i of argv
else
return my formatErrorMessage("Argument Error", "--model requires a model name parameter", "validation")
end if
else if arg is "--provider" then
if i < argCount then
set i to i + 1
set requestedProvider to item i of argv
if requestedProvider is not "auto" and requestedProvider is not "ollama" and requestedProvider is not "claude" then
return my formatErrorMessage("Argument Error", "--provider must be 'auto', 'ollama', or 'claude'", "validation")
end if
else
return my formatErrorMessage("Argument Error", "--provider requires a provider name parameter", "validation")
end if
else if arg is "--format" then
if i < argCount then
set i to i + 1
set outputFormat to item i of argv
if outputFormat is not "png" and outputFormat is not "jpg" and outputFormat is not "pdf" then
return my formatErrorMessage("Argument Error", "--format must be 'png', 'jpg', or 'pdf'", "validation")
end if
else
return my formatErrorMessage("Argument Error", "--format requires a format parameter", "validation")
end if
-- Handle boolean flags
else if arg is "--fullscreen" or arg is "-f" then
set forceFullscreen to true
else if arg is "--window" or arg is "-w" then
set captureMode to "window"
else if arg is "--multi" or arg is "-m" then
set multiWindow to true
else if arg is "--verbose" or arg is "-v" then
set verboseLogging to true
else if arg is "--quiet" or arg is "-q" then
set quietMode to true
-- Handle positional argument (output path for old-style compatibility)
else if not (arg starts with "-") and command is "capture" and not outputSpecified then
set outputPath to arg
set outputSpecified to true
end if
set i to i + 1
end repeat
-- Handle analyze command
if command is "analyze" then
-- For analyze command, appIdentifier contains the image path
return my analyzeImageWithAI(appIdentifier, analysisQuestion, visionModel, requestedProvider)
end if
-- For capture command, determine capture mode
if captureMode is "" then
if forceFullscreen or appIdentifier is "" then
set captureMode to "screen"
else
-- App specified, default to window capture
set captureMode to "window"
end if
end if
-- Set default output path if none provided
if outputPath is "" then
set timestamp to do shell script "date +%Y%m%d_%H%M%S"
-- Create model-friendly filename with app name
if appIdentifier is "" or appIdentifier is "fullscreen" then
set appNameForFile to "fullscreen"
else
set appNameForFile to my sanitizeAppName(appIdentifier)
end if
-- Determine extension based on format
set fileExt to outputFormat
if fileExt is "" then set fileExt to defaultScreenshotFormat
set outputPath to "/tmp/peepit_" & appNameForFile & "_" & timestamp & "." & fileExt
else
-- Check if user specified a directory for multi-window mode
if multiWindow and outputPath ends with "/" then
set timestamp to do shell script "date +%Y%m%d_%H%M%S"
set appNameForFile to my sanitizeAppName(appIdentifier)
set fileExt to outputFormat
if fileExt is "" then set fileExt to defaultScreenshotFormat
set outputPath to outputPath & "peepit_" & appNameForFile & "_" & timestamp & "." & fileExt
else if outputFormat is not "" and not (outputPath ends with ("." & outputFormat)) then
-- Apply format if specified but not in path
set outputPath to outputPath & "." & outputFormat
end if
end if
-- Validate output path
if outputSpecified and not my isValidPath(outputPath) then
return my formatErrorMessage("Argument Error", "Output path must be an absolute path starting with '/'.", "validation")
end if
-- Resolve app identifier with detailed diagnostics
if appIdentifier is "" or appIdentifier is "fullscreen" then
set appInfo to {appName:"fullscreen", bundleID:"fullscreen", isRunning:true, resolvedBy:"fullscreen"}
else
set appInfo to my resolveAppIdentifier(appIdentifier)
end if
if appInfo is missing value then
set errorDetails to "Could not resolve app identifier '" & appIdentifier & "'."
-- Provide specific guidance based on identifier type
if appIdentifier contains "." then
set errorDetails to errorDetails & " This appears to be a bundle ID. Common issues:" & linefeed
set errorDetails to errorDetails & "• Bundle ID may be incorrect (try 'com.apple.' prefix for system apps)" & linefeed
set errorDetails to errorDetails & "• App may not be installed" & linefeed
set errorDetails to errorDetails & "• Use 'osascript peepit.scpt list' to see available apps"
else
set errorDetails to errorDetails & " This appears to be an app name. Common issues:" & linefeed
set errorDetails to errorDetails & "• App name may be incorrect (case-sensitive)" & linefeed
set errorDetails to errorDetails & "• App may not be installed or running" & linefeed
set errorDetails to errorDetails & "• Try the full app name (e.g., 'Activity Monitor' not 'Activity')" & linefeed
set errorDetails to errorDetails & "• Use 'osascript peepit.scpt list' to see running apps"
end if
return my formatErrorMessage("App Resolution Error", errorDetails, "app resolution")
end if
set resolvedAppName to appName of appInfo
set resolvedBy to resolvedBy of appInfo
my logVerbose("App resolved: " & resolvedAppName & " (method: " & resolvedBy & ")")
-- Bring app to front
set frontError to my bringAppToFront(appInfo)
if frontError is not "" then return frontError
-- Smart multi-window detection for AI analysis
if analyzeMode and resolvedAppName is not "fullscreen" and not forceFullscreen then
-- Check how many windows the app has
set windowStatus to my getAppWindowStatus(resolvedAppName)
try
set statusClass to class of windowStatus
if statusClass is not text and statusClass is not string then
-- It's a success record
set totalWindows to totalWindows of windowStatus
if totalWindows > 1 and not multiWindow and captureMode is not "screen" then
-- Automatically enable multi-window mode for AI analysis
set multiWindow to true
my logVerbose("Auto-enabling multi-window mode for AI analysis (app has " & totalWindows & " windows)")
end if
end if
on error
-- Continue without auto-enabling
end try
end if
-- Pre-capture window validation for better error messages
if (multiWindow or captureMode is "window") and resolvedAppName is not "fullscreen" then
set windowStatus to my getAppWindowStatus(resolvedAppName)
-- Check if it's an error (string starting with prefix) or success (record)
try
set statusClass to class of windowStatus
if statusClass is text or statusClass is string then
-- It's an error message
if multiWindow then
set contextError to "Multi-window capture failed: " & windowStatus
set contextError to contextError & linefeed & "💡 Suggestion: Try basic screenshot mode without --multi flag"
else
set contextError to "Window capture failed: " & windowStatus
set contextError to contextError & linefeed & "💡 Suggestion: Try full-screen capture mode without --window flag"
end if
return contextError
else
-- It's a success record
set statusMsg to message of windowStatus
my logVerbose("Window validation passed: " & statusMsg)
end if
on error
-- Fallback if type check fails
my logVerbose("Window validation status check bypassed")
end try
end if
-- Handle multi-window capture
if multiWindow then
set capturedFiles to my captureMultipleWindows(resolvedAppName, outputPath)
-- Check if it's an error (string) or success (list)
try
set capturedClass to class of capturedFiles
if capturedClass is text or capturedClass is string then
return capturedFiles -- Error message
end if
on error
-- Continue with list processing
end try
-- If AI analysis requested, analyze all captured windows
if analyzeMode and (count of capturedFiles) > 0 then
set analysisResults to {}
set allSuccess to true
repeat with fileInfo in capturedFiles
set filePath to item 1 of fileInfo
set windowTitle to item 2 of fileInfo
set windowIndex to item 3 of fileInfo
set analysisResult to my analyzeImageWithAI(filePath, analysisQuestion, visionModel, requestedProvider)
if analysisResult starts with scriptInfoPrefix and analysisResult contains "Analysis Complete" then
-- Extract just the answer part from the analysis
set answerStart to (offset of "💬 Answer:" in analysisResult) + 10
set answerEnd to (offset of (scriptInfoPrefix & "Analysis via") in analysisResult) - 1
if answerStart > 10 and answerEnd > answerStart then
set windowAnswer to text answerStart thru answerEnd of analysisResult
else
set windowAnswer to analysisResult
end if
set end of analysisResults to {windowTitle:windowTitle, windowIndex:windowIndex, answer:windowAnswer, success:true}
else
set allSuccess to false
set end of analysisResults to {windowTitle:windowTitle, windowIndex:windowIndex, answer:analysisResult, success:false}
end if
end repeat
-- Format multi-window AI analysis results
return my formatMultiWindowAnalysis(capturedFiles, analysisResults, resolvedAppName, analysisQuestion, visionModel, quietMode)
else
-- Process successful capture without AI
return my formatMultiOutput(capturedFiles, resolvedAppName, quietMode)
end if
else
-- Single capture
set screenshotResult to my captureScreenshot(outputPath, captureMode, resolvedAppName)
if screenshotResult starts with scriptInfoPrefix then
return screenshotResult -- Error message
else
set modeDescription to "full screen"
if captureMode is "window" then set modeDescription to "front window only"
-- If AI analysis requested, analyze the screenshot
if analyzeMode then
set analysisResult to my analyzeImageWithAI(screenshotResult, analysisQuestion, visionModel, requestedProvider)
if analysisResult starts with scriptInfoPrefix and analysisResult contains "Analysis Complete" then
-- Successful analysis
return analysisResult
else
-- Analysis failed, return screenshot success + analysis error
return scriptInfoPrefix & "Screenshot captured successfully! 📸" & linefeed & "• File: " & screenshotResult & linefeed & "• App: " & resolvedAppName & linefeed & "• Mode: " & modeDescription & linefeed & linefeed & "⚠️ AI Analysis failed:" & linefeed & analysisResult
end if
else
-- Regular screenshot without analysis
return my formatCaptureOutput(screenshotResult, resolvedAppName, modeDescription, quietMode)
end if
end if
end if
on error generalErrorMsg number generalErrorNum
if appSpecificErrorOccurred then error generalErrorMsg number generalErrorNum
return my formatErrorMessage("Execution Error", generalErrorMsg, "error " & generalErrorNum)
end try
end run
--#endregion Main Script Logic (on run)
--#region Usage Function
on usageText()
set LF to linefeed
set scriptName to "peepit.scpt"
set outText to "PeepIt v1.0.0 - Screenshot automation that actually works! 👀 → 📸 → 💾" & LF & LF
set outText to outText & "USAGE:" & LF
set outText to outText & " peepit [app] [options] # Screenshot app or fullscreen" & LF
set outText to outText & " peepit analyze <image> \"question\" [opts] # Analyze existing image" & LF
set outText to outText & " peepit list # List running apps" & LF
set outText to outText & " peepit help # Show this help" & LF & LF
set outText to outText & "COMMANDS:" & LF
set outText to outText & " [app] App name or bundle ID (optional, defaults to fullscreen)" & LF
set outText to outText & " analyze Analyze existing image with AI vision" & LF
set outText to outText & " list, ls List all running apps with window info" & LF
set outText to outText & " help, -h Show this help message" & LF & LF
set outText to outText & "OPTIONS:" & LF
set outText to outText & " -o, --output <path> Output file or directory path" & LF
set outText to outText & " -f, --fullscreen Force fullscreen capture" & LF
set outText to outText & " -w, --window Single window capture (default with app)" & LF
set outText to outText & " -m, --multi Capture all app windows separately" & LF
set outText to outText & " -a, --ask \"question\" AI analysis of screenshot" & LF
set outText to outText & " --model <model> AI model (e.g., llava:7b)" & LF
set outText to outText & " --provider <provider> AI provider: auto|ollama|claude" & LF
set outText to outText & " --format <fmt> Output format: png|jpg|pdf" & LF
set outText to outText & " -v, --verbose Enable debug output" & LF
set outText to outText & " -q, --quiet Minimal output (just file path)" & LF & LF
set outText to outText & "EXAMPLES:" & LF
set outText to outText & " # Basic captures" & LF
set outText to outText & " peepit # Fullscreen" & LF
set outText to outText & " peepit Safari # Safari window" & LF
set outText to outText & " peepit Safari -o ~/Desktop/safari.png # Specific path" & LF
set outText to outText & " peepit -f -o screenshot.jpg --format jpg # Fullscreen as JPG" & LF & LF
set outText to outText & " # Multi-window capture" & LF
set outText to outText & " peepit Chrome -m # All Chrome windows" & LF
set outText to outText & " peepit Safari -m -o ~/screenshots/ # To directory" & LF & LF
set outText to outText & " # AI analysis" & LF
set outText to outText & " peepit Safari -a \"What's on this page?\" # Screenshot + analyze" & LF
set outText to outText & " peepit -f -a \"Any errors visible?\" # Fullscreen + analyze" & LF
set outText to outText & " peepit analyze photo.png \"What is this?\" # Analyze existing" & LF
set outText to outText & " peepit Terminal -a \"Show the error\" --model llava:13b" & LF & LF
set outText to outText & " # Other commands" & LF
set outText to outText & " peepit list # Show running apps" & LF
set outText to outText & " peepit help # This help" & LF & LF
set outText to outText & "Note: When using with osascript, quote arguments and escape as needed:" & LF
set outText to outText & " osascript peepit.scpt Safari -a \"What's shown?\"" & LF & LF
set outText to outText & "AI Analysis Features:" & LF
set outText to outText & " • Smart provider detection: auto-detects Ollama or Claude CLI" & LF
set outText to outText & " • Smart multi-window: Automatically analyzes ALL windows for multi-window apps" & LF
set outText to outText & " - App has 3 windows? Analyzes all 3 and reports on each" & LF
set outText to outText & " - Use -w flag to force single window analysis" & LF
set outText to outText & " • Ollama: Local inference with vision models (recommended)" & LF
set outText to outText & " - Supports direct image file analysis" & LF
set outText to outText & " - Priority: qwen2.5vl:7b > llava:7b > llava-phi3:3.8b > minicpm-v:8b" & LF
set outText to outText & " • Claude: Limited support (CLI doesn't analyze image files)" & LF
set outText to outText & " - Claude CLI detected but can't process image files directly" & LF
set outText to outText & " - Use Ollama for automated image analysis" & LF
set outText to outText & " • One-step: Screenshot + analysis in single command" & LF
set outText to outText & " • Two-step: Analyze existing images separately" & LF
set outText to outText & " • Timeout protection: 90-second timeout prevents hanging" & LF & LF
set outText to outText & "Multi-Window Features:" & LF
set outText to outText & " • --multi creates separate files with descriptive names" & LF
set outText to outText & " • Window titles are sanitized for safe filenames" & LF
set outText to outText & " • Files named as: basename_window_N_title.ext" & LF
set outText to outText & " • Each window is focused before capture for accuracy" & LF & LF
set outText to outText & "Notes:" & LF
set outText to outText & " • Default behavior: App specified = window capture, No app = full screen" & LF
set outText to outText & " • Requires Screen Recording permission in System Preferences" & LF
set outText to outText & " • Accessibility permission may be needed for window enumeration" & LF
set outText to outText & " • Window titles longer than " & maxWindowTitleLength & " characters are truncated" & LF
set outText to outText & " • Default capture delay: " & (captureDelay as string) & " second(s) (optimized for speed)" & LF
return outText
end usageText
--#endregion Usage Function
