Bot configuration
Every setting on the Aimogen Pro chatbot configuration and API parameter screens, with what each one actually controls.
AI Chatbot › Bot Configuration and AI Chatbot › API Parameters.
Identity#
| Setting | What it controls |
|---|---|
Bot Name | Displayed name, and how the bot refers to itself. Stored as ai_message_preppend |
Bot Description | A short role description, shown in the header on some designs. Stored as ai_role |
Bot Context | The system prompt. Who the bot is, what it knows, what it must not do |
Bot Initial Messages (One Per Line) | Opening messages. Several lines means one is chosen at random |
Bot Avatar Image | Avatar URL |
User Name | How the visitor is labelled |
User Avatar Image | Default visitor avatar |
Use User Avatar If A Logged In User Using The Chatbot | Use the WordPress avatar for logged-in visitors |
Bot Context#
This single field determines most of the bot behaviour. It supports placeholders, which makes it dynamic:
You are the assistant for %%blog_title%%.
The visitor is %%user_display_name%% (role: %%user_role%%).
They are currently reading: %%post_title%%
Page content: %%post_content_plain_text%%
Answer questions about this page. If the question is not about it, say so and
suggest they use the site search.Useful placeholders here: %%post_title%%, %%post_content%%, %%post_content_plain_text%%,
%%post_excerpt%%, %%post_cats%%, %%post_tags%%, %%post_link%%, %%user_name%%,
%%user_display_name%%, %%user_firstname%%, %%user_lastname%%, %%user_email%%, %%user_role%%,
%%user_id%%, %%user_language%%, %%blog_title%%, %%author_name%%, %%current_date_time%%,
%%privacy_policy_url%%.
Model and capabilities#
| Setting | What it controls |
|---|---|
Bot Model | Which model answers. Latency matters more than eloquence here |
Bot Assistant Name | Use an AI Assistant instead. Doing so disables personas |
Enable Bot Vision | Visitors can send images. Requires a vision-capable model |
Enable File Uploads To The Bot | General file uploads. See Files and PDF chat |
Enable Chatbot Instant Responses | Reply from a cache or a scripted answer where possible, without a model call |
Allow Users To Stop Response Creation When Streaming | A stop button during streaming |
Default Chat Mode | text or image — whether the bot writes or generates images by default |
API parameters#
AI Chatbot › API Parameters
| Setting | Default | What it controls |
|---|---|---|
Chatbot Max Token Count | — | Output cap per reply. Keep it modest; long chat replies are rarely wanted |
Chatbot Temperature | 1 | Lower is more consistent, higher more varied. For support bots, 0.3–0.7 |
Chatbot Top_p | 1 | Adjust this or temperature, not both |
Presence Penalty | 0 | Discourages returning to topics already covered |
Frequency Penalty | 0 | Discourages repeating words |
Store AI Prompts On OpenAI's Part | — | Whether the provider may retain the request |
Conversation context#
| Setting | What it controls |
|---|---|
Max Chat Messages To Send As API Context | How many previous turns are resent with each message |
Max Chat Context Size (Characters) | A character cap on that history |
Single Message Max Input Length (Characters) | Longest message a visitor may send |
Disable Sending Of Empty Chat Messages | Rejects empty submissions |
Prompt modifiers#
| Setting | What it controls |
|---|---|
Prepend Text To All Textual AI Prompts From AI Chatbots | Added before every message |
Append Text To All Textual AI Prompts From AI Chatbots | Added after every message |
Conversation Starters | Suggested questions shown to the visitor |
Prompt Templates | Predefined prompts the visitor can select |
Prompts Editable By Users | Whether visitors can edit a selected template |
Text For Prompt Templates Selection | The label on the selector |
The append field is a good place for a persistent formatting instruction, such as "always answer in at most three short paragraphs".
Response processing#
| Setting | What it controls |
|---|---|
Convert Markdown To HTML In AI Generated Content | Renders markdown. Usually wanted |
Enable Katex (Mathematical Expressions Rendering) | Renders LaTeX maths |
Enable Chatbot Full HTML Response Processing | Allows HTML in replies |
Strip JavaScript From Chatbot HTML Responses | Removes scripts. Leave this on |
Disable Chatbot Modern Response Processing | Falls back to the older rendering path |
Chatbot Response Delay (ms) | Artificial delay before the reply appears |
Typewriter Delay (ms) | Speed of the typing effect |
Interaction#
| Setting | What it controls |
|---|---|
Enable Message Copying By Clicking It | Click to copy a message |
Enable Chat Message Editing By Clicking It | Visitors can edit and resend |
Scroll To Bottom Of The Form On New Messages | Auto-scroll |
Up/Down Keys Show Older Messages In Chat | Shell-style history |
Download Chat Conversation To File As | Export format for the download button |
Retry behaviour#
Chat End of Conversation Retry Count (Settings › General Settings) retries when the model produces a
reply the plugin reads as ending the conversation. Each retry is another paid request — 1 or 2 is enough.
Related#
- Personas
- Appearance and user interface
- Knowledge and context
- Shortcode reference — the per-instance attribute equivalents
Still stuck? Open a support ticket and include the diagnostics from Aimogen Pro › System & Logs › System Info.