451 words
2 minutes
AI Auto-Response Settings Guide
What is this page?
The /prompt page is a settings panel where you manage the four rules that govern how AI auto-responds to donor chats. Each section has an independent toggle whose state is saved to the server the moment you flip it.
Idle Auto Reply
- When no chats arrive for a while, the AI gathers the most recent messages and proactively says something on its own.
- Idle threshold (15–3600 sec, default 15) — must elapse since the last chat before the AI speaks.
- Collect count (0–50, default 5; 0 = always speak) — number of recent chats fed to the AI as context.
- Cooldown (15–3600 sec, default 15) sets the minimum gap between auto-replies, and the custom prompt (max 2000 chars) lets you dictate tone.
Priority Capture of Donor Chats
- Right after a donation, this filter promotes the donor’s own most valuable chat from a short buffer window into the priority queue.
- Buffer window (5–60 sec, default 6) — duration after a donation during which chats are collected for comparison.
- Donor window (1–168 hours, default 24) — how long a recent donation keeps a viewer recognised as a donor.
- Pick count (1–20, default 2) — how many chats from the buffer end up queued.
Amount-Based Chat Filter
- Limits responses to viewers whose donations exceed a threshold. Each platform has its own minimum.
- Choose a base platform and the other 9 platforms are auto-filled by exchange rates. Editing a field manually locks it as a manual override.
- Supported platforms: YouTube · Twitch · SOOP · Kick · CHZZK · TikTok · Douyu · Bilibili · Huya · Douyin.
0in any field means “no limit”, and an exchange summary appears when you save.
Donor Chat Memory (Premium)
- Stores N past messages per donor so the AI can reuse them as context. The multi-rule card system lets you apply different policies per tier.
- Each card holds: reference platform, minimum donation, donor window (1–168h), chat retention (1–168h), and max messages (5–200).
- Stored conversations are included in the AI auto-answer prompt when that donor chats again, producing context-aware replies.
- Use “Add rule” to create up to 20 cards; ✕ on a card removes it. After saving, the live counter shows tracked donors and stored messages.
Auto-Response Use Cases
- Unmanned mode: combine IAR + Auto Answer so the AI keeps the chat alive even while you’re AFK.
- VIP touch: with DCM, the AI remembers regulars’ nicknames and interests for warmer replies.
- Quiet down noise: tighten DCF Amount so non-donor chats are ignored and credits/tokens last longer.
- Event tone: drop event context like “Today is the OO event” into IAR’s custom prompt and even idle replies adopt the event tone.
When changes are not reflected in the URL
Double-click the source with the URL in OBS, check the ‘Refresh browser when scene becomes active’ checkbox, and click the ‘Refresh cache of current page’ button.
Share Article
If this article helped you, please share it with others!
AI Auto-Response Settings Guide
https://blog.donastar.tv/en/posts/prompt-guide/Related PostsSmart
1
AI Auto Answer Guide
Donastar Streaming ToolsAI detects donor messages and delivers responses tuned to the streamer's tone to an OBS widget (and the 'AI Auto Answer monitor'), with optional TTS voice. It…
2
AI VTuber Guide
Donastar Streaming ToolsLinked to AI Auto Answer, a Live2D character appears on screen and reacts with expressions, motions, and lip-sync whenever the AI responds. You get a…
3
Donor Chat Memory Guide
Donastar Streaming Tools/chat-memory is where you browse, search, and summarise the per-donor message histories collected via DCM. It uses a two-pane layout: details on the left…
4
Smart Donation Image Generation Guide
Donastar Streaming ToolsAI generates images in real time from prompts your viewers type in chat and shows them on your stream widget.
5
Smart Voice Chat User Guide
Donastar Streaming ToolsA feature that reads chat in real-time and automatically generates voice responses. It collects chat from various platforms, generates text responses, then…
Random PostsRandom



