OffgridMobile Test Flows - Comprehensive Inventory This document catalogs every testable flow in the application.
Coverage Legend:
✅ = Fully tested (Unit + Integration/RNTL + E2E)
⚠️ = Partially tested (missing some layers)📝 = Planned but not implemented
❌ = No test coverage
Current Overall Coverage: ~90% of P0 flows, ~65% of all flows
See docs/test/TEST_PRIORITY_MAP.md for detailed test status and docs/test/TEST_SPEC_FORMAT.md for how to write tests.
1. Onboarding & First Launch
#
Flow
Description
1.1
Fresh install onboarding
App opens → welcome screens → permissions → complete
1.2
Onboarding with model download
Onboarding → recommended model → download → ready to chat
1.3
Skip onboarding download
Skip model download → home with no model
1.4
Onboarding persistence
Partially complete → kill app → resumes where left off
1.5
Onboarding completed flag
Completed once → never shown again
2. Authentication & Security
#
Flow
Description
2.1
First-time passphrase setup
Settings → Security → set passphrase → enabled
2.2
Passphrase confirmation match
Enter passphrase twice → must match
2.3
Passphrase confirmation mismatch
Entries differ → error shown
2.4
Passphrase strength indicator
Weak/medium/strong feedback
2.5
App lock on background
App backgrounded → returns → lock screen
2.6
Successful unlock
Correct passphrase → app unlocks
2.7
Failed unlock attempt
Wrong passphrase → error + attempt counted
2.8
Multiple failed attempts
Track count across attempts
2.9
Lockout triggered
5 failures → 5 minute lockout
2.10
Lockout countdown
Shows remaining time
2.11
Lockout expiry
Timer ends → can try again
2.12
Lockout persists across app restart
Locked → kill app → still locked
2.13
Change passphrase
Requires old passphrase → set new
2.14
Disable passphrase
Requires passphrase → then disabled
2.15
Auth state persistence
Enabled state survives app restart
#
Flow
Description
3.1
Microphone permission request
First voice input → permission prompt
3.2
Microphone permission granted
Permission given → recording works
3.3
Microphone permission denied
Denied → graceful error message
3.4
Camera permission request
First image attachment → permission prompt
3.5
Camera permission granted
Permission given → camera works
3.6
Camera permission denied
Denied → graceful error, can still use gallery
3.7
Storage permission (Android)
Required for downloads on older Android
3.8
Permission settings redirect
"Go to settings" → opens app settings
#
Flow
Description
4.1
Home with no model
Shows "download a model" prompt
4.2
Home with model not loaded
Shows model name + "tap to load"
4.3
Home with model loaded
Shows model name + memory usage
4.4
Model loading from home
Tap model → loading indicator → loaded
4.5
Quick model selector
Open selector → pick model → loads
4.6
Memory usage display
Shows RAM used / available
4.7
Recent conversations
Shows last N chats
4.8
New conversation from home
"New Chat" → creates + navigates
4.9
Tap recent conversation
Navigates to that chat
4.10
Device info on home
Shows device model + capabilities
4.11
Home refresh
Pull to refresh → updates model list
#
Flow
Description
5.1
Models tab loads
Shows list of available models from HuggingFace
5.2
Model list loading state
Spinner while fetching
5.3
Model list error state
API error → retry button
5.4
Search models by name
Type query → list filters
5.5
Search clears
Clear search → full list
5.6
Filter by credibility - LM Studio
Toggle → only LM Studio models
5.7
Filter by credibility - Official
Toggle → only official models
5.8
Filter by credibility - Verified
Toggle → only verified quantizers
5.9
Filter by model type - Vision
Toggle → only vision-capable models
5.10
Multiple filters combined
LM Studio + Vision → intersection
5.11
Clear all filters
Reset → full list
5.12
Empty filter results
No matches → "no models found"
5.13
Pull to refresh
Refresh model list from API
5.14
Pagination / infinite scroll
Load more models on scroll
#
Flow
Description
5.15
View model card
Tap model → expanded details
5.16
View available files
List of quantization options
5.17
File size display
Shows size in GB/MB
5.18
Quantization explanation
Info about Q4_K_M etc
5.19
Vision model indicator
Badge for vision-capable
5.20
Author display
Shows who uploaded
5.21
Credibility badge
LM Studio/Official/Verified badge
5.22
Model recommendations
"Recommended for your device"
#
Flow
Description
5.23
Start foreground download
Select file → download starts
5.24
Download progress display
Progress bar + percentage
5.25
Download speed display
MB/s indicator
5.26
Download ETA
Estimated time remaining
5.27
Download pause
Pause button → pauses
5.28
Download resume
Resume → continues from where left off
5.29
Download cancel
Cancel → partial file cleaned up
5.30
Download complete
Progress → 100% → model in list
5.31
Download while app backgrounded
Continues in background (foreground mode)
5.32
Multiple concurrent downloads
Download 2+ models at once
5.33
Download queue
Start multiple → queued
5.34
Insufficient storage
Not enough space → error before download
5.35
Download network error
Network drops → error + retry option
5.36
Download resumes after network
Network back → can resume
5.37
Vision model download
Downloads main + mmproj file
5.38
Vision model partial download
Main file done, mmproj fails → handles gracefully
5.39
Duplicate download prevention
Already downloaded → shows "downloaded"
5.40
Already downloading prevention
In progress → can't start again
#
Flow
Description
5.41
View downloaded models
Tab/section shows downloaded
5.42
Downloaded model info
Size on disk, date downloaded
5.43
Delete downloaded model
Delete → removed from disk + list
5.44
Delete confirmation
Confirm dialog before delete
5.45
Delete active model
Warning that it will be unloaded
5.46
Model file corruption
Corrupted file detected → error + delete option
5.47
Scan for untracked models
Find models added externally
5.48
Model recovery
App killed → model files still exist → recovered
6. Image Model Management
#
Flow
Description
6.1
Image models tab
Switch to image models list
6.2
Browse available image models
List from HuggingFace repos
6.3
Image model details
Size, style, backend info
6.4
Download image model
Download → extract → ready
6.5
Image model extraction progress
Zip extraction progress
6.6
View downloaded image models
List downloaded SD models
6.7
Delete image model
Remove from disk
6.8
Backend indicator
MNN vs QNN vs Core ML badge
6.9
QNN compatibility
Shows if device supports QNN
6.10
Image model size warning
Large model → warning
6.11
Core ML model browsing (iOS)
Fetch available models from Apple repos
6.12
Core ML model caching
Cache model list with 5-min TTL
6.13
Core ML partial failure
Some repos fail → return partial results
7. Background Downloads (Cross-Platform)
#
Flow
Description
7.1
Start background download
Select → uses DownloadManager (Android) / URLSession (iOS)
7.2
Notification appears
System notification shows progress
7.3
Download manager screen
View all background downloads
7.4
Progress in download manager
Percentage + bytes downloaded
7.5
Pause background download
Pause → paused state in notification
7.6
Resume background download
Resume → continues
7.7
Cancel background download
Cancel → removed
7.8
Download completes
Notification → complete
7.9
App killed during download
Download continues via system
7.10
App reopens - download complete
Detects completion → moves file → adds to list
7.11
App reopens - download in progress
Shows current progress
7.12
App reopens - download failed
Shows failed state
7.13
Clear completed downloads
Remove from list
7.14
Multiple background downloads
Several simultaneous
7.15
iOS download parity
Same DownloadManagerModule interface via URLSession
7.16
iOS polling compatibility
startProgressPolling/stopProgressPolling are no-ops on iOS
7.17
iOS completed download localUri
Completed download includes localUri (moved from temp)
8. Model Loading & Memory
#
Flow
Description
8.1
Load text model
Select model → loading → loaded
8.2
Loading progress indicator
Shows loading state
8.3
Loading time display
Shows how long loading took
8.4
Model loaded confirmation
Success message/indicator
8.5
Load failure - corrupt file
Error message + suggestion
8.6
Load failure - insufficient memory
OOM error → suggestion to try smaller
8.7
Load failure - unknown error
Generic error + retry
8.8
Unload model
Unload → memory freed
8.9
Unload confirmation
Confirm before unloading
8.10
Memory freed display
Shows reclaimed memory
8.11
Switch text models
A loaded → select B → A unloads → B loads
8.12
Switch with active generation
Warn that generation will stop
8.13
Load image model
Select → loading → ready
8.14
Image model loading progress
Shows loading state
8.15
Switch image models
Swap image models
8.16
Memory warning threshold
Low memory → warning before load
8.17
Memory critical
Very low → refuse to load
8.18
Performance mode
Model stays loaded between generations
8.19
Memory mode
Model unloads after generation
8.20
GPU layer loading
Load N layers to GPU
8.21
GPU loading failure
GPU unavailable → falls back to CPU
8.22
Model session caching
Same system prompt → cached session
#
Flow
Description
Status
9.1
Send text message
Type → send → message appears
✅ E2E: 02-text-generation.yaml
9.2
Message appears in chat
User message shown immediately
✅
9.3
Generation starts
Loading/thinking indicator
✅
9.4
Streaming tokens
Tokens appear one by one
✅
9.5
Streaming performance
Smooth token display
✅ E2E validates
9.6
Generation completes
Full response shown
✅
9.7
Response saved
Response persists after app restart
✅
9.8
Time to first token
TTFT displayed
⚠️ RNTL only
9.9
Tokens per second
Generation speed shown
⚠️ RNTL only
9.10
Total token count
Shows prompt + response tokens
⚠️ RNTL only
#
Flow
Description
Status
9.11
Stop generation
Stop button → generation halts
✅ E2E: 03-stop-generation.yaml
9.12
Partial response saved
Stopped response is kept
✅
9.13
Retry generation
Retry → regenerates last response
⚠️ Unit + RNTL only
9.14
Edit user message
Edit → changes saved
⚠️ Unit + RNTL only
9.15
Regenerate from edit
Edit message → regenerates from that point
⚠️ Unit only
9.16
Delete messages after
Delete from point → later messages removed
⚠️ Unit + RNTL only
#
Flow
Description
9.17
Conversation history
Previous messages in context
9.18
System prompt applied
Project system prompt used
9.19
Context length limit
Old messages dropped when limit hit
9.20
Context truncation
Smart truncation preserves recent
9.21
Long conversation
50+ messages → still works
9.22
Empty conversation
First message has no history
Generation Settings Impact
#
Flow
Description
9.23
Temperature low (0.1)
Deterministic output
9.24
Temperature high (1.5)
Random/creative output
9.25
Max tokens reached
Response stops at limit
9.26
Top-p affects output
Different sampling
9.27
Repeat penalty works
Reduces repetition
9.28
Thread count affects speed
More threads = faster
#
Flow
Description
9.29
Generation error
Error message shown
9.30
Model crashes mid-generation
Handles gracefully
9.31
OOM during generation
Error + suggestion
9.32
No model loaded
Prompt to load model
#
Flow
Description
9.33
Thinking blocks
<think> tags parsed + displayed separately
9.34
Code blocks
Markdown code blocks rendered
9.35
Markdown rendering
Bold, italic, lists, etc.
9.36
Long response
Handles multi-paragraph responses
9.37
Unicode/emoji in response
Rendered correctly
#
Flow
Description
9.38
Show generation details on
Details visible in message
9.39
Show generation details off
Details hidden
9.40
GPU indicator
Shows if GPU used
9.41
Model name in metadata
Which model generated
10. Intent Classification
#
Flow
Description
10.1
Clear text intent
"What is X?" → text
10.2
Clear image intent
"Draw a cat" → image
10.3
Question patterns → text
"How do I...", "Why is...", etc.
10.4
Generation patterns → image
"Generate", "create", "paint", etc.
10.5
Art style patterns → image
"Oil painting of", "anime style"
10.6
Code patterns → text
"Write a function", "debug this"
10.7
SD-specific → image
"Stable diffusion", "negative prompt"
10.8
Ambiguous prompt - pattern mode
Falls back to text
10.9
Ambiguous prompt - LLM mode
Uses model to classify
10.10
LLM classification loading
Shows "analyzing" state
10.11
Classifier model switch
Uses different model for classification
10.12
Classifier model swap back
Returns to main model after
10.13
Intent cache hit
Repeated prompt uses cache
10.14
Intent cache size limit
Cache eviction when full
10.15
Quick check for UI hints
Hint shown before sending
10.16
Pattern false positive
"Show me how to draw" → text not image
10.17
Pattern false negative
Edge case not caught → correct fallback
#
Flow
Description
11.1
Auto-detect triggers generation
"Draw X" → image generated
11.2
Manual mode generation
Force image → any prompt works
11.3
Image model loads
Model loads if not loaded
11.4
Generation progress
Step X/Y shown
11.5
Preview during generation
Partial image shown
11.6
Preview updates
Updates as steps complete
11.7
Generation completes
Final image shown
11.8
Image in chat
Image appears in conversation
11.9
Image in gallery
Image added to gallery
11.10
Image saved to disk
File persists
#
Flow
Description
11.11
Cancel image generation
Cancel → stops mid-generation
11.12
Partial image on cancel
Shows what was generated
11.13
Retry image generation
Retry → regenerates
11.14
Different seed
Retry produces different image
#
Flow
Description
11.15
Negative prompt
Specify exclusions → applied
11.16
Custom steps (low)
10 steps → fast but rough
11.17
Custom steps (high)
30+ steps → slow but detailed
11.18
Guidance scale low
More creative interpretation
11.19
Guidance scale high
Strict prompt following
11.20
Custom resolution square
512x512
11.21
Custom resolution landscape
768x512
11.22
Custom resolution portrait
512x768
11.23
Resolution not divisible by 8
Rejected or auto-adjusted
11.24
Thread count affects speed
More threads = faster
#
Flow
Description
11.25
MNN backend (Android)
CPU-based generation
11.26
QNN backend (Android)
NPU-accelerated generation
11.27
QNN not available
Falls back to MNN
11.28
Auto backend selection
Picks best for device
11.29a
Core ML backend (iOS)
ANE-accelerated generation via ml-stable-diffusion
#
Flow
Description
11.30
Generation error
Error message shown
11.31
No image model loaded
Prompt to download/select
11.32
OOM during generation
Error + lower resolution suggestion
11.33a
Platform routing
Platform.select routes to correct native module
#
Flow
Description
11.33
Metadata saved
Prompt, steps, seed, etc. stored
11.34
Metadata displayed
View generation settings
11.35
Model name in metadata
Which SD model used
12. Vision Models (Image Understanding)
#
Flow
Description
12.1
Load vision model
Model + mmproj both loaded
12.2
Attach image from gallery
Pick image → attached
12.3
Attach image from camera
Take photo → attached
12.4
Image attachment preview
Thumbnail shown in input
12.5
Remove attachment
X button → removed
12.6
Send with image
Message + image sent
12.7
Image in context
Model "sees" the image
12.8
Generate about image
"What's in this?" → description
12.9
Multiple images
Several images in conversation
12.10
Image without vision model
Warning that model can't see images
12.11
Large image handling
Resize/compress if needed
12.12
Image format support
JPG, PNG, etc.
#
Flow
Description
13.1
Attach text document
Pick .txt → content extracted
13.2
Attach code file
Pick .py/.js → content extracted
13.3
Document preview
Shows filename + snippet
13.4
Document in context
Content included in prompt
13.5
Summarize document
"Summarize this" → summary
13.6
Large document
Truncated to fit context
13.7
Unsupported format
Error message
13.8
Multiple documents
Several files attached
13.9
Document + image
Both attached together
#
Flow
Description
14.1
Tap mic to start
Recording begins
14.2
Recording indicator
Visual feedback while recording
14.3
Recording duration
Shows time elapsed
14.4
Stop recording
Stop → transcription starts
14.5
Transcription progress
Loading indicator
14.6
Transcription complete
Text appears in input
14.7
Partial transcription
Live updates while speaking
14.8
Cancel recording
Cancel → discarded
14.9
Empty recording
Very short → ignored
14.10
Long recording
Handles extended speech
14.11
No Whisper model
Prompt to download
14.12
Whisper model loading
Load before first transcription
14.13
Transcription accuracy
Quality matches model size
14.14
Background noise
Handles noisy environment
14.15
Multiple languages
Works with selected language
#
Flow
Description
14.16
View Whisper models
List tiny/small/base/medium
14.17
Download Whisper model
Download → ready
14.18
Switch Whisper model
Change to different size
14.19
Delete Whisper model
Remove downloaded model
14.20
English-only model
Smaller, English only
14.21
Multilingual model
Larger, many languages
14.22
Model recommendation
Based on device RAM
#
Flow
Description
15.1
Create new conversation
New button → empty chat
15.2
Conversation created timestamp
Created date stored
15.3
Auto-generate title
First message → title derived
15.4
Edit conversation title
Manual title change
15.5
Conversation updated timestamp
Updates on new message
15.6
Delete conversation
Delete → removed
15.7
Delete confirmation
Confirm dialog
15.8
Delete removes messages
All messages deleted
15.9
Delete removes images
Associated gallery images removed
#
Flow
Description
15.10
View conversations list
List of all chats
15.11
Conversations sorted by date
Most recent first
15.12
Conversation preview
Shows last message snippet
15.13
Switch conversations
Tap → switches to chat
15.14
Empty conversations list
"No conversations" state
15.15
Swipe to delete
Swipe action deletes
15.16
Search conversations
Search by title/content
#
Flow
Description
15.17
Conversations persist
Survives app restart
15.18
Messages persist
All messages saved
15.19
Conversation state restored
Active conversation remembered
15.20
Scroll position restored
Returns to same position
Conversation with Project
#
Flow
Description
15.21
Start from project
Project → new chat → system prompt applied
15.22
Project linked
Conversation shows project
15.23
Project deleted
Conversations keep their system prompt
16. Chat UI & Interactions
#
Flow
Description
16.1
User message styling
Right-aligned, colored
16.2
Assistant message styling
Left-aligned, different color
16.3
System message styling
Distinct style
16.4
Message timestamp
Shows when sent
16.5
Message grouping
Adjacent same-role grouped
16.6
Streaming message style
Different while streaming
16.7
Long message display
Handles paragraphs correctly
16.8
Code syntax highlighting
Code blocks colored
16.9
Link detection
URLs tappable
16.10
Copy message
Long press → copy
16.11
Message actions menu
Edit, delete, retry options
#
Flow
Description
16.12
Type message
Text input works
16.13
Multiline input
Enter creates new line
16.14
Send button enabled
Enabled when text present
16.15
Send button disabled
Disabled when empty
16.16
Send button disabled - no model
Disabled without model
16.17
Send button disabled - generating
Disabled during generation
16.18
Clear input after send
Input clears on send
16.19
Input character limit
Handles very long input
16.20
Paste long text
Paste works correctly
16.21
Unicode/emoji input
Handles special characters
16.22
Whitespace-only blocked
Can't send empty/spaces
16.23
Keyboard handling
Keyboard pushes up input
16.24
Input focus
Tap input → keyboard opens
#
Flow
Description
16.25
Auto-scroll new message
Scrolls to bottom on new message
16.26
Auto-scroll during stream
Follows streaming content
16.27
Manual scroll up
Can scroll to old messages
16.28
Manual scroll - no auto-scroll
Doesn't force to bottom
16.29
Scroll to bottom button
Tap → scrolls to bottom
16.30
Pull to refresh
Refresh conversation?
#
Flow
Description
16.31
Empty chat state
"Send a message to start"
16.32
Loading model state
Shows loading indicator
16.33
Generating state
Shows generating indicator
16.34
Error state
Shows error message
16.35
Offline state
Shows offline indicator
#
Flow
Description
16.36
Attachment button
Opens attachment options
16.37
Attachment options
Camera, gallery, document
16.38
Attachment preview
Shows thumbnail/name
16.39
Remove attachment
X removes it
16.40
Multiple attachments
Several shown
16.41
Image mode toggle
Switch to image generation mode
16.42
Image mode indicator
Shows when in image mode
17. Projects (System Prompt Presets)
#
Flow
Description
17.1
View projects list
List of all projects
17.2
Default projects
4 default projects present
17.3
Create new project
New → form → save
17.4
Project name required
Can't save without name
17.5
Project description
Optional description
17.6
Project system prompt
Required system prompt
17.7
Project icon selection
Choose icon
17.8
Save project
Save → appears in list
17.9
Edit project
Edit existing → save changes
17.10
Delete project
Delete → removed
17.11
Delete confirmation
Confirm before delete
17.12
Duplicate project
Copy with new name
#
Flow
Description
17.13
Start chat from project
Project → new chat → prompt applied
17.14
Project preview
See system prompt preview
17.15
Project linked conversations
View chats using this project
17.16
Projects persist
Survive app restart
#
Flow
Description
18.1
View gallery
Grid of all generated images
18.2
Gallery loading
Loading indicator
18.3
Empty gallery
"No images" state
18.4
Image thumbnails
Thumbnails load correctly
18.5
Image grid layout
Responsive grid
18.6
Scroll gallery
Scroll through many images
18.7
Gallery sorted by date
Most recent first
#
Flow
Description
18.8
Filter by conversation
Select conversation → filtered
18.9
Clear filter
Show all images
18.10
Empty filter result
No images for conversation
#
Flow
Description
18.11
Tap image
Opens fullscreen viewer
18.12
Fullscreen viewer
Image fills screen
18.13
Pinch to zoom
Zoom in/out
18.14
Pan zoomed image
Move around
18.15
Close fullscreen
Back → returns to gallery
18.16
View image metadata
See prompt, settings
18.17
Share image
Share to other apps
18.18
Save to device gallery
Export to photos
#
Flow
Description
18.19
Delete single image
Delete → removed
18.20
Delete confirmation
Confirm before delete
18.21
Multi-select mode
Enter selection mode
18.22
Select multiple images
Tap to select/deselect
18.23
Select all
Select all visible
18.24
Delete selected
Delete multiple at once
18.25
Clear gallery
Delete all images
18.26
Deletion removes file
File deleted from disk
#
Flow
Description
18.27
Gallery persists
Images survive restart
18.28
Metadata persists
Prompt/settings survive
18.29
Orphan file handling
File exists but not in DB
#
Flow
Description
19.1
Open settings
Settings tab → screen opens
19.2
Settings sections
Grouped by category
19.3
Navigate to model settings
Tap → model settings screen
19.4
Navigate to voice settings
Tap → voice settings screen
19.5
Navigate to security
Tap → security screen
19.6
Navigate to storage
Tap → storage screen
19.7
Navigate to device info
Tap → device info screen
19.8
Back navigation
Back → returns to settings
#
Flow
Description
19.9
Temperature slider
Adjust 0.0 - 2.0
19.10
Temperature value display
Shows current value
19.11
Max tokens input
Set max response length
19.12
Max tokens validation
Positive integer only
19.13
Top-p slider
Adjust 0.0 - 1.0
19.14
Repeat penalty slider
Adjust 1.0 - 2.0
19.15
Context length input
Set context window
19.16
Context length validation
Reasonable range
19.17
Thread count slider
1 to device max
19.18
Batch size input
Set batch size
19.19
GPU toggle
Enable/disable GPU
19.20
GPU layers slider
0 to max layers
19.21
GPU requires reload notice
"Requires model reload"
19.22
Loading strategy toggle
Performance vs memory
19.23
Reset to defaults
Reset all model settings
19.24
Settings saved immediately
Changes auto-save
Image Generation Settings
#
Flow
Description
19.25
Steps slider
Adjust 1 - 50
19.26
Guidance scale slider
Adjust 1.0 - 20.0
19.27
Width input
Set image width
19.28
Height input
Set image height
19.29
Resolution validation
Divisible by 8
19.30
Image threads slider
1 to device max
19.31
Image settings apply next gen
Takes effect on next generation
Intent Detection Settings
#
Flow
Description
19.32
Image generation mode
Auto vs manual
19.33
Auto-detect method
Pattern vs LLM
19.34
Classifier model select
Choose classifier model
19.35
Show generation details
Toggle metadata display
#
Flow
Description
19.36
All settings persist
Survive app restart
19.37
Settings sync
Changes reflect immediately
19.38
Settings migration
Old settings preserved on update
#
Flow
Description
20.1
View total storage used
Sum of all models
20.2
Text models storage
Size of text models
20.3
Image models storage
Size of image models
20.4
Whisper models storage
Size of Whisper models
20.5
Gallery storage
Size of generated images
20.6
Available storage
Free disk space
20.7
Storage warning
Low storage warning
20.8
Per-model sizes
Size of each model
20.9
Sort by size
Largest first
20.10
Delete from storage screen
Remove model from here
20.11
Bulk delete
Select multiple → delete
#
Flow
Description
21.1
View device info
Screen opens
21.2
Device model
Shows phone model
21.3
OS version
Shows Android/iOS version
21.4
Total RAM
Shows device RAM
21.5
Available RAM
Shows free RAM
21.6
Total storage
Shows device storage
21.7
Available storage
Shows free storage
21.8
Emulator detection
Flags if running in emulator
21.9
GPU info
Shows GPU capabilities
21.10
Model recommendations
Suggests models for device
21.11
Refresh device info
Update readings
22. Download Manager Screen
#
Flow
Description
22.1
View download manager
Screen opens
22.2
Active downloads list
Shows in-progress downloads
22.3
Download progress
Percentage and bytes
22.4
Download status
Running/paused/completed/failed
22.5
Pause download
Pause button works
22.6
Resume download
Resume button works
22.7
Cancel download
Cancel button works
22.8
Completed downloads
Shows completed with checkmark
22.9
Failed downloads
Shows error state
22.10
Retry failed
Retry button works
22.11
Clear completed
Remove from list
22.12
Empty state
No active downloads message
23. App Lifecycle & Persistence
#
Flow
Description
23.1
App to background
State preserved
23.2
Return from background
Resumes correctly
23.3
Background during generation
Generation continues or pauses
23.4
App force killed
State saved
23.5
Reopen after kill
State restored
23.6
Active model remembered
Model ID persisted
23.7
Active conversation remembered
Returns to same chat
23.8
Settings restored
All settings restored
23.9
Downloads restored
Background downloads recovered
23.10
Generation interrupted
Partial response handled
23.11
First launch after update
Migration runs
23.12
AsyncStorage corruption
Graceful handling
#
Flow
Description
24.1
Bottom tab navigation
Home/Chat/Models/Gallery/Settings
24.2
Tab state preserved
Tab remembers state
24.3
Stack navigation
Push/pop screens
24.4
Back button (Android)
Hardware back works
24.5
Gesture navigation
Swipe back works
24.6
Deep nested back
Multiple levels → returns correctly
24.7
Navigate during generation
Warning or handled gracefully
24.8
Navigate during download
Download continues
#
Flow
Description
25.1
No internet - model browse
Error message + retry
25.2
No internet - download
Error + retry option
25.3
No internet - HuggingFace API
Graceful error
25.4
Network timeout
Timeout handled
25.5
Network slow
Progress still shows
25.6
Network recovery
Auto-retry or manual
25.7
Partial download resume
Resume from where left off
#
Flow
Description
25.8
Corrupt model file
Detected + error shown
25.9
Model load OOM
Error + smaller model suggestion
25.10
Model incompatible
Version mismatch error
25.11
Model missing
Deleted externally → removed from list
#
Flow
Description
25.12
Generation crash
Error message
25.13
Generation OOM
Error + suggestion
25.14
Generation timeout
Long generation handled
25.15
Native module crash
Graceful recovery
#
Flow
Description
25.16
Storage full
Error before download
25.17
Storage full mid-download
Handles gracefully
25.18
File system error
Generic error handling
25.19
Permission denied
Error message
#
Flow
Description
25.20
LocalDream unavailable (iOS)
"Not supported" message
25.21
Background download unavailable (iOS)
Falls back to foreground
25.22
GPU unavailable
Falls back to CPU
25.23
QNN unavailable
Falls back to MNN
26. Edge Cases & Stress Tests
#
Flow
Description
26.1
Very long message
10000+ characters
26.2
Message with only spaces
Blocked/ignored
26.3
Message with only newlines
Handled
26.4
Unicode edge cases
Emoji, RTL, etc.
26.5
Special characters
<, >, &, etc.
26.6
Code injection attempts
Handled safely
26.7
Extremely long word
No spaces → handled
#
Flow
Description
26.8
Very long conversation
100+ messages
26.9
Many conversations
50+ conversations
26.10
Many downloaded models
20+ models
26.11
Large gallery
500+ images
26.12
Rapid message sending
Spam send button
26.13
Rapid model switching
Quick switch back and forth
#
Flow
Description
26.14
Download during generation
Both work
26.15
Multiple downloads
All progress
26.16
Generation + voice recording
Both work
26.17
Switch conversation during generation
Warning or handled
26.18
Delete model during generation
Warning
26.19
App background during generation
Generation pauses/continues
#
Flow
Description
26.20
Low memory warning
System warns → app responds
26.21
System kills background app
State persisted
26.22
Memory pressure during generation
Handles gracefully
Total flows: 360+
Category
Count
Onboarding
5
Authentication
15
Permissions
8
Home Screen
11
Text Model Management
48
Image Model Management
13
Background Downloads
17
Model Loading
22
Text Generation
41
Intent Classification
17
Image Generation
36
Vision Models
12
Document Attachments
9
Voice Input
22
Conversations
23
Chat UI
42
Projects
16
Gallery
29
Settings
38
Storage Settings
11
Device Info
11
Download Manager
12
App Lifecycle
12
Navigation
8
Error Handling
23
Edge Cases
22