Documentation

System Requirements

Before installing EchoAI, ensure your server meets the following requirements.

Installation Guide

Follow these steps to get EchoAI up and running on your Sngine site. Each step below includes the most common thing that trips people up, based on real support tickets.

AI Chat New

The core of EchoAI – a powerful, real‑time conversational AI experience.

Key Features

• Real‑time streaming – responses appear character by character (DeepSeek).
• Thinking Mode – enables DeepSeek's reasoning engine, showing live reasoning steps with a timer.
• Multi‑language responses – auto‑detects the user's language and replies accordingly.
• Code block minimap – VS‑Code‑style minimap for long code snippets.
• Image generation – integrated with OpenAI DALL‑E or DeepSeek (text description).
• File & image uploads – attach a photo, PDF, or document via the paperclip button. The file stages as a small preview above the message box, so you can add a caption (or send it as-is) before it ever posts — nothing is sent the moment you pick a file. Sending an image with a caption gets a real AI reply, the same as a normal message.
• Stop Generating – cancel a response mid-stream with one click.
• Time-aware responses – the AI knows the current date and time, so it can answer things like "what day is it" or reason about recency correctly.
• Time-zone conversion New – EchoAI detects each user's browser time zone and now actually uses it: ask "what time is 15:30 UTC for me?" and get a real, computed local time back – not a guess.
• Scope Restriction New – admins can confine the assistant to a specific set of topics (e.g. a single game or product) using a real AI classification call rather than keyword blocking. See AI Management.
• Hero / Unit Combos New – admin-curated pairing data grounds tactical/combo questions in verified answers instead of an improvised one. See AI Management.
• Troop / Unit Calculator New – admins enter real per-level resource costs and players get exact training/upgrade calculations computed on the server, never guessed by the AI. See AI Management.
• Upload Chats Coming Soon – import conversations from other platforms.

Thinking Mode

When Thinking Mode is enabled (per package permission), the AI shows its reasoning process before delivering the final answer. This is perfect for complex queries where transparency builds trust.

Token & Pricing System

EchoAI includes a flexible, monetizable token system that gives you full control over pricing.

Pricing Models

Package Management

Admins can create and manage token packages from the admin panel → Packages. Each package includes:

• Tokens – the initial token balance
• Daily / Weekly / Monthly Limits – usage caps per period
• Reset Period – daily, weekly, or monthly token reset
• Thinking Mode – enable/disable for this package
• Memory Limit – max memory entries per user

Memory System

EchoAI remembers important facts about users, creating a personalised experience across conversations.

Features

• Per‑conversation memory – AI remembers facts within a specific chat.
• Global memory – if enabled, the AI remembers across all conversations.
• Memory limits – per‑user, per‑entry, and per‑conversation limits.
• Package‑based memory limit – premium packages can have more memory entries.
• Import / Export – admins can export memories as JSON.

Admin Controls

• Clear all memories
• Clear memories for a specific user
• View memory usage statistics

Personality System

Allow users to customise their AI experience – available to selected packages.

What Users Can Customise

• AI Name – give the AI a custom name
• Color Theme – choose from 9 themes, including animated colours
• Background URL – set a custom background image
• Disable Filters – bypass the blocklist (if allowed by admin)
• Custom JSON Responses – trigger‑based custom replies
• URL Mask – generate a custom URL for your AI persona (e.g., yourdomain.com/ai/yourname)
• Quick Action Buttons – add up to 4 custom buttons that trigger predefined prompts or actions, and change or replace any button at any time.

Image Generation

Users can generate images directly from the chat interface using natural language prompts.

• OpenAI DALL‑E – full image generation with local caching.
• DeepSeek – generates a detailed text description of the requested image (useful for budget‑friendly setups).
• Local Storage – images are saved to disk in organised year/month/day folders and served from your own server rather than a third-party link.
• Duplicate Caching – identical prompts return the cached image, saving tokens.
• Push notification on completion – if OneSignal is configured, the user gets a browser notification the moment their image is ready, even if they've navigated away from the chat — see Push Notifications.

Admins can set the cost per image in Settings → Costs & Tokens.

Folder Settings

Organise your conversations like never before. EchoAI now lets users create custom folders to group chats, assign colours and emojis, and save conversations for quick access.

How It Works

• Click the "+" button in the sidebar to create a new folder.
• Enter a folder name, select a colour from the palette, and choose an emoji.
• Once created, the folder appears in your sidebar. Drag any conversation into it, or use the conversation menu to move it.
• Folders are saved per user – each user has their own set of folders.
• Admins can manage folder settings globally, including limiting the number of folders per user, and set files/chats/folders to auto-delete after a chosen period (e.g. 1 week).

AI Media Search

Supercharge your community by enabling AI-powered search for Sngine posts directly from the EchoAI interface.

How It Works

• When enabled, a icon appears in the chat input area.
• Click the icon to activate Search Mode – the input placeholder changes to "Search public posts...".
• Type your query (e.g., "recent updates about AI") and send. The AI will search Sngine's public posts.
• The response will include a list of matching posts with:
• Post content (snippet)
• Author – who posted it
• URL – a clickable link to the original post
• Users can then click the link to view the full post on your site.

Library

Now all images and files are automatically organised in a central Library, with a refreshed design — pill-style tabs for switching between views and clearer, more polished cards for every file and image.

When a user generates an image or sends a file, it is automatically added to their personal library. The library offers three views:

• All – shows both images and files together.
• Images – only generated or uploaded images.
• Files – only uploaded files (PDF, DOC, etc.).
• Playground – saved Coding Playground projects, if the plugin is installed.

Users can browse, sort, and delete items from their library. This makes it easy to keep track of all generated content and uploaded documents in one place.

Push Notifications New

EchoAI can send a real browser push notification when something the user asked for finishes in the background — most commonly, "Your image is ready" once an AI image generation completes. Tapping the notification takes the user straight back into that conversation, even if they've closed the tab.

Push notifications are powered by OneSignal, a free push notification service. You'll need a (free) OneSignal account to use this feature.

Enabling / Disabling the "Image Ready" Notification

On the same Notifications settings page, toggle "Notify user when an AI-generated image is ready". This is on by default once OneSignal itself is enabled.

Extensions New

EchoAI now supports a powerful Extensions system that lets you expand functionality with plugins and themes.

Navigate to AdminCP of EchoAI → Modules → Extensions to manage everything. From there, you can:

• View – see all installed extensions and their status.
• Get Support – access documentation and support for each extension.
• Buy – purchase premium extensions from the marketplace.
• Install – upload a ZIP file via drag‑and‑drop and it will be installed automatically.
• Pause – temporarily disable an extension without uninstalling.
• Delete – permanently remove an extension.

The system handles both plugins and themes. Users can download a ZIP file from the marketplace or developer, then simply drag and drop it into the Extensions area – the installation is fully automatic.

🤖 EchoBot (Premium Plugin)

Transform your Sngine chat into a powerful command center! With EchoBot, users can effortlessly manage their entire social presence using natural language – no more navigating complex menus.

Buy link: https://portasale.com/items/echobot-command-center-for-sngine-powered-by-echoai/1146

Autonomous Bots New

Beyond running commands for a real logged-in user, EchoBot can now connect to a Sngine Fake Generator account and turn it into a fully autonomous, AI-powered community member — posting, commenting, reacting, messaging, sending/accepting friend requests, and running its own Pages/Groups/Stories entirely on its own, on a schedule you control.

• Per-bot schedule – active days, active hours, and hourly/daily action & AI-request limits, so activity always stays within bounds you set.
• Presence – realistic online/offline cycling, or an "always on" mode.
• Ignore / block list – tell a bot to stop responding to specific users, the same way a real person could.
• Shared-hosting friendly – runs via a self-triggered scheduler that piggybacks on normal site traffic, so no shell/cron access is required; a real crontab entry is supported too if your host offers one, for more consistent timing.
• AI disclosure – every connected bot account carries a persistent marker in its own profile bio (and, optionally, a short tag on the content it generates) so it's always identifiable as AI-operated. This is a deliberate design choice, not a limitation.
• Full AdminCP – a dashboard of every connected bot, a per-bot settings editor, an activity log, a built-in connection/permission test, and a safe uninstall path.

Commands that EchoBot can run:

Content Generation Commands (AI writes, then asks where to publish):

• write something about [topic]
• generate a post about [topic]
• generate content about [topic]
• write about [topic]
• write a blog about [topic] (if you don't specify where to publish)
• generate a post about [topic] with an image

Publishing Target Commands (after generating content):

• profile
• page [page name]
• group [group name]
• post to [page name] page
• post to [group name] group

Friend Commands:

• send friend request to [name]
• add [name] as friend
• accept friend request from [name]
• accept [name]'s friend request
• decline friend request from [name]
• remove [name] from friends
• unfriend [name]

Advanced Commands (specify where to publish directly):

• post [text] to my profile
• post [text] to page [name]
• post [text] to group [name]
• post a photo of [subject] to [page name] page
• write a blog about [topic] on page [name]
• write a blog about [topic] on group [name]
• post [text] to [user name]'s wall

Confirmation Responses:

• yes / y / confirm / do it / proceed / go ahead – confirm delete
• no / n / cancel / stop / nevermind / never mind – cancel delete

💻 Coding Playground (Free)

A powerful AI coding assistant with live preview, multi-file support, project management, and advanced code generation — all inside the Playground.

Giving ability for users to generate code, full structure code to HTML/CSS/JS, PHP, Python and more.

You can find the addon on Extensions → Plugin → coding-playground

Dashboard

The home base for admins – a quick overview of your entire system.

• Total Users – registered users count
• Total Tokens – tokens in circulation
• Text Generations – total AI text requests
• Images Generated – total images created
• AI Commands – number of custom commands
• Blocked Topics – active blocklist entries

Quick action buttons give you one‑click access to Users, Packages, Commands, Blocklist, and Settings.

Users

The user management hub – view, search, and manage all registered accounts.

Actions Available

• Search & Filter – by name, email, status (active/blocked), token balance
• Token Management – add, remove, or set exact token balance
• Block / Unblock – with custom reason
• Clear Memory – remove all AI memories for a user
• Delete Chats – permanently remove all conversations
• View Stats – message count, conversation count, daily/weekly/monthly usage, language preference, memory mode

Packages

Manage your pricing and service tiers – both Sngine subscriptions and EchoAI's own standalone token packages.

Sngine Packages

View and edit AI‑specific settings for each Sngine subscription package:

• Token amount
• Daily / Weekly / Monthly / Hourly limits
• Reset period
• Cost per text, image, file
• Thinking Mode toggle
• Memory limit

EchoAI Standalone Subscriptions New

As of Update 1.3, EchoAI packages no longer have to ride on top of Sngine's own subscription system — you can give any EchoAI package its own price and sell it directly through EchoAI's built-in Payments system (PayPal or the custom bank-transfer provider), completely independent of Sngine.

• Set a Price on any package to make it purchasable directly.
• Link a package to a PayPal Plan ID for automatic recurring billing (see Payments → PayPal for how to create one).
• Leave the price at 0 to keep a package free/manual, exactly as before — nothing changes for packages you don't touch.
• Tokens, limits, reset period, Thinking Mode toggle, and Memory limit all work the same as Sngine-linked packages.
• Bulk token management: refresh, pause, resume, delete.

Payments New

As of Update 1.3, EchoAI can accept payment for its own standalone packages directly — without routing through Sngine's subscription system at all. Two providers are supported, and you can enable either one, both, or neither.

AI Management New

Fine‑tune the AI's behaviour and capabilities.

Commands

Create custom slash commands (e.g., /help, /summarize) that users can type to trigger specific AI responses. Each command can have its own system prompt.

Blocklist

Add forbidden words or phrases. If a user's message contains a blocked word, the request is rejected with a custom response.

API Keys

Manage your DeepSeek and OpenAI API keys, including the new GPT-5.6 model family (Sol, Terra, and Luna versions). Test each key with a single click to verify it's working.

Scope Restriction New

The opposite of the Blocklist: instead of listing what's forbidden, list exactly what the assistant is allowed to talk about (e.g. a specific game, product, or service). Turn on Strict Scope Mode and add one or more allowed topics – every message is then classified against that list by a real AI call before it's answered, not just keyword matching, so ordinary phrasing that never says the topic's name by name still gets recognised correctly. Greetings and questions about the assistant itself always stay allowed. If no API key is configured yet, or the classifier call fails for any reason, Scope Restriction fails open – it never locks users out due to a transient error.

Hero / Unit Combos New

Built for gaming communities. Enter a combo (a set of unit or hero names, plus your own description of how they work together), and whenever a user's message mentions any of those names, that combo's description is automatically pulled into the AI's context as verified data – so it answers with your team's actual, tested pairing instead of an AI-invented one that only sounds plausible.

Troop / Unit Calculator New

Also built for gaming communities that manage resource-based progression (training or upgrading units/troops using in-game resources). Enter the real cost to train or upgrade each level (up to 4 configurable resource types – not locked to any specific game's terminology), and a player can ask something like "I have 1M mana, 800k ore, 500k wood, and 2M stone – how many level V troops can I train?" and get an exact number back. The arithmetic is always computed by EchoAI's own code from your configured costs – never left to the AI to calculate – so the answer is exact every time, including which resource is the limiting factor.

Content

Tools for communicating with your users and collecting feedback.

Announcements

Send global news, maintenance alerts, or feature updates to all users. Announcements appear as in‑app notifications.

Reports & Feedback

View user‑submitted reports and feedback. Mark them as resolved, reply, or take action (e.g., ban a user).

System

Core configuration and technical administration.

Settings

The central configuration panel. Adjust everything from the AI name and avatar to pricing models, memory limits, and feature toggles.

• General – AI brand name, avatar icon, system prompt, default language
• AI Configuration – default API providers, model selection, temperature
• Costs & Tokens – pricing model, token costs, thinking mode cost
• Memory – context window size, max memory entries
• Language – allow users to change language
• Personality – enable personality system, select allowed packages
• Feedback – disable like/dislike, copy, or report buttons
• Social Post Search – enable/disable AI search for Sngine posts
• Notifications – configure OneSignal push notifications, see Push Notifications

Live Stats

Real‑time monitoring: online users, today's activity, token usage trends, device breakdown, top users, and feedback statistics.

Logs

Historical record of admin actions – every settings change, user block, token adjustment, and more, logged with IP and user agent for security auditing. Update 1.3 adds a dedicated Error Log alongside the admin action log, so you can see server-side errors (failed API calls, database issues, etc.) without needing shell/FTP access to your host's raw PHP error log.

Cron Jobs

EchoAI relies on a daily cron job to reset user tokens and usage counters. This is essential for the token system to work correctly.

The Cron Command

Add this to your server's crontab (e.g., via HestiaCP, cPanel, or crontab -e):

0 0 * * * php /home/user/web/yourdomain.com/public_html/echoai/cron/reset_tokens.php >> /home/user/logs/echoai_reset.log 2>&1

What It Does

• Checks every user's last_reset timestamp against their package's reset_period (daily, weekly, or monthly).
• If the period has passed, it resets the token balance to the package's token amount and sets usage counters to 0.
• Updates last_reset to the current time.

Other Cron Jobs

If you've enabled EchoAI's standalone Payments system, also add the payment cron job — it checks pending PayPal subscriptions and handles renewal bookkeeping between webhook calls:

*/30 * * * * php /home/user/web/yourdomain.com/public_html/echoai/cron/payment_cron.php >> /home/user/logs/echoai_payment.log 2>&1

Testing a Cron Job

You can run any of these scripts manually to test them:

php /home/user/web/yourdomain.com/public_html/echoai/cron/reset_tokens.php

If you see output like Completed. X users had their tokens reset., it's working.

Updating EchoAI

Keeping EchoAI up‑to‑date ensures you have the latest features, bug fixes, and security patches. This guide covers the manual update process – an automatic updater is in development.

📦 Manual Update Process (Recommended)

This is the only method currently available – it gives you full control and is the safest way to update.

1. Download the latest version – Obtain the new EchoAI ZIP package (e.g., EchoAI_v1.4.zip) from the marketplace or your purchase email.
2. Extract the Update folder – Inside the ZIP, locate the Update folder. This folder contains all files that have changed since the previous version — it mirrors the exact folder structure of your live install, so every file already tells you where it belongs.
3. Check for SQL updates – Inside the Update folder, look for one or more .sql files. Note: not every update includes one — some updates only modify PHP files. If a SQL file is present, proceed to step 4; otherwise, skip to step 5. These SQL files are written to be safe to run more than once (they check whether a column/table already exists before adding it), so re-running one by mistake won't break anything.
4. Apply the SQL changes – Log in to your database management tool (e.g., phpMyAdmin), select your EchoAI database, open the SQL tab, and paste in the contents of each .sql file (or use the Import feature). This adds any new tables or columns required for the update's features.
5. Replace the updated files – Copy the contents of the Update folder to your server's EchoAI installation directory (e.g., /public_html/echoai/), preserving the folder structure. Important: Do not delete or overwrite entire folders — only copy the individual files from the Update folder into their corresponding locations. For example, if the Update folder contains admin/settings.php, replace only that one file, not your entire admin/ folder. Replacing entire folders can corrupt your installation, wipe custom edits, or cause missing-file errors.
6. Clear every layer of cache – this is the single most common reason an update "doesn't seem to have installed":
   • Browser cache — hard refresh with Ctrl+Shift+R (Windows/Linux) or Cmd+Shift+R (Mac).
   • Server-side PHP OPcache — many hosts cache compiled PHP and won't pick up a changed file instantly. Restart PHP-FPM if you have shell access, or use your host's "Restart PHP" / "Select PHP Version" control panel button, which typically forces a reset.
   • CDN or full-page cache — if you use Cloudflare, LiteSpeed Cache, or similar, purge the cache for your domain.

Note: An automatic update feature is currently in development and will be available in future releases. For now, please use the manual process described above.

🔄 Post‑Update Verification Checklist

Don't just assume an update landed — confirm it, in this order:

1. Confirm the files actually changed — open one of the updated files directly on the server (via File Manager) and check the content matches what you expect, rather than trusting the upload alone.
2. Hard refresh and re-test the specific feature the update was about, not just "does the chat still load."
3. Check Admin → Error Logs for anything new right after the update — this surfaces problems (like a missing SQL step) far faster than waiting for a user to report them.
4. Verify the token system and cron jobs still run correctly — an update that touches the database occasionally needs the cron job to run once before everything settles.
5. If something still looks wrong after all of the above, check the FAQ or contact support with the exact update version and what you're seeing.

Changelog

All notable changes to EchoAI are documented here.

Frequently Asked Questions

Quick answers to the most common questions about EchoAI.