Appearance
Theme
White Contrast
Paper
Midnight
Graphite
OLED
Custom
Speech bubbles
Solid
Soft gradient
Living aurora
Organic drift
Gradient modes gently animate each user's bubble colour. Text contrast is recalculated across every gradient stop, and motion automatically stops when Reduced Motion is enabled.
Message composer border
Theme default
Slow glow
Rotating gradient
Rainbow orbit
Fade in / out
Dashed
Dotted
Blinking amber
High-vis amber
Select on the left and preview the exact resting composer treatment on the right. Animated choices automatically become static when the operating system requests Reduced Motion.
Custom theme name
✦ Lucky Gen Lucky ✓ Save named theme Save Export JSON
Import JSON
Delete saved
↺ Reset draft Reset
UI font
Load installed fonts
Experimental AI
AI model
Choose the model for normal /ai text chat. Hetzner and Local Llama use server-managed inference; OpenAI models require your own key.
Extended OpenAI jobs (>10 min)
Off keeps background responses privacy-first with store:false. Enable only when you want long jobs to remain retrievable beyond OpenAI's short background polling window.
Paid output limit
1,024 tokens · very short
2,048 tokens · economy / may truncate
4,096 tokens
8,192 tokens
16,384 tokens · long-answer default
32,768 tokens
65,536 tokens · high cost
128,000 tokens · model maximum / very high cost
Confirm paid GPT-6 Astra requests
One /ai command creates at most one paid OpenAI response. The output cap includes visible and reasoning tokens; input tokens are billed separately.
OpenAI wallet: no calls tracked on this device yet.
Experimental image AI
Off
On request
Automatic description
Per-device setting · checking experimental image AI…
Image AI timeout
sec
Per-device setting · 15–300 seconds · default 180 seconds. Saved automatically and applied to each image-analysis request.
Recent context
When enabled, recent text from the active channel is sent to the selected /ai text provider so follow-up questions can use the conversation.
YouTube transcript summaries: checking server support…
Prompt history
When enabled, Arrow Up/Down in an empty or /ai composer recalls recent AI prompts stored locally for this profile so you can edit and resend them.
Builder mode
When enabled, /ai returns a self-contained HTML app in a sandboxed rendered overlay instead of printing HTML into chat. You can also use /ai build without enabling the preference.
Builder thread
When on, /ai sent while a Builder preview is open updates that exact build using only its prior Builder requests and latest HTML.
Open AI Slop Gallery
Media
Read .txt attachments in chat
Off by default. When enabled on this device, Rantlist fetches only a bounded plain-text excerpt for .txt attachments. The text is never interpreted as HTML, Markdown or JavaScript; full text opens separately as sandboxed text/plain.
Ask video quality before upload
Maximum video file per upload
64 MiB · default
128 MiB
256 MiB
512 MiB · 0.5 GiB
The administrator must enable larger uploads. Your storage quota and available server capacity are checked separately.
Image and ordinary file upload maximum: 16 MiB (set by the administrator).
Off by default. When enabled, sending an ordinary video from Video, File, Camera, recording, drag/drop or recovered native sharing asks for an HLS transcoding preset before upload bytes are sent. The original master is always retained.
Enable /yt video downloads
Off by default. When enabled, type /yt followed by a public media URL. X/Twitter clips use the best quality that fits the server/storage ceiling automatically; YouTube, Odysee and other site-specific yt-dlp extractors show the qualities actually reported for that item.
Default yt-dlp quality
Preserve source
AAC/M4A · 256 kbps
AAC/M4A · 192 kbps
AAC/M4A · 128 kbps
AAC/M4A · 96 kbps
AAC/M4A · 64 kbps · data saver
Streamable HLS · 256 kbps
Streamable HLS · 192 kbps
Streamable HLS · 128 kbps
Streamable HLS · 64 kbps · data saver
When a URL-only message is supported by a site-specific yt-dlp extractor, Rantlist opens the import chooser with file and streamable HLS qualities, estimated size, current storage quota and long-media confirmation. HLS choices publish a stream-only index.m3u8 for chat and Stage instead of retaining the downloaded source file. Unsupported URLs post normally. The default above only preselects a quality; every import still requires confirmation.
Media previews
Click to reveal
Show immediately
Channel media viewer
Stories / Shorts
Classic inline
Stories / Shorts is the default: tap any channel image or video to enter the full-screen vertical media feed. Choose Classic inline only if you want videos to play inside message bubbles.
Speech
Audio/video transcripts
Off
On request
Automatic
Transcription language
Voice memo limit
30 seconds
1 minute
3 minutes
5 minutes
10 minutes
15 minutes
20 minutes
30 minutes
Per-device recording limit · default 15 minutes · server maximum 30 minutes.
Read aloud
Reading speed
Slow
Normal
Fast
Very fast
System voice
History
Load newest
Nothing
25 items
50 items
100 items
200 items
500 items
Clear visible room history