How wpForo AI protects your credits with caching, rate limits, feature toggles, and smart optimizations
1 Feature Toggles
Each AI feature can be independently enabled or disabled in Dashboard → wpForo → Settings → AI Features. Disabled features consume zero credits. Please find all AI Settings in this page >>
Semantic Search
Core vector search — finds semantically similar topics
Search Enhancement
AI-generated summary & recommendations on search results
Topic Suggestions
Suggests similar/related topics while creating a new topic
Topic Summarization
AI-generated summary of long topic discussions
Translation
Multi-language post translation
AI Chatbot
Forum assistant chatbot (also configurable by user group)
Content Moderation
Spam, toxicity, and compliance detection — each toggleable separately
Image Indexing
Process images in posts for visual search (costs +1 credit per image)
Example: Search-Only Configuration
Enable only “Semantic Search” and optionally “Search Enhancement.” Disable everything else. Your credits will only be spent on content indexing and search queries — nothing else.
2 Intelligent Caching
wpForo AI caches results at multiple levels. When a cached result is available, zero credits are consumed and rate limits are bypassed.
| Cache Type | Duration | How It Saves Credits |
|---|---|---|
| Translation | Permanent | Each post is translated once per language and cached forever. Repeat views cost nothing. |
| Search Results | 24 hours | Same search query returns cached result. Popular searches cost 1 credit for the first user, 0 for everyone else that day. |
| Search Enhancement | 24 hours | AI summary and recommendations cached per query. Regenerated only when search results change. |
| Topic Summary | 24 hours | Cached per topic + quality level + language. Regenerated only when the topic receives new replies. |
| Tenant Status | 5 minutes | Subscription and credit info cached to avoid API calls on every page load. |
| Analytics | 1 hour | Dashboard analytics data cached to prevent frequent queries. |
How cache keys work
Cache entries are invalidated automatically when the underlying data changes. For example, a topic summary cache key includes the reply count and last modification time — so when someone replies, a fresh summary is generated. But if nobody replies, the cached version is served for free.
3 Rate Limiting
Rate limits prevent excessive usage and protect your credit balance from abuse, runaway scripts, or high-traffic spikes.
Per-User Daily Limits (Frontend)
Configurable per feature in the admin settings:
- Guests — Limited daily searches, translations, and suggestions (tracked by IP address)
- Logged-in users — Configurable daily limit per feature (tracked by user ID)
- Moderators & admins — Exempt from daily limits
Per-Tenant Rate Limits (Backend)
In addition to frontend daily limits, the backend enforces per-minute request limits based on your subscription plan. Higher plans receive higher limits, ensuring fair usage across all tenants while preventing any single site from consuming excessive resources.
Cache + Rate Limits Work Together
Rate limits are checked after the cache. If a cached result exists, neither credits nor rate limit quota are consumed. A popular search query that 100 users run in one day costs exactly 1 credit.
4 Quality Tiers & Model Selection
Every AI-powered feature lets you choose a quality tier. Lower tiers use smaller, faster, and cheaper AI models:
| Quality Tier | Speed | Cost | Best For |
|---|---|---|---|
| Fast | Fastest | Lowest | High-traffic forums, budget-conscious setups |
| Balanced | Fast | Low | General use, good quality/cost ratio |
| Advanced | Moderate | Medium | Complex queries, professional forums |
| Premium | Moderate | Highest | Maximum quality, nuanced responses |
Quality tiers are available for: Search Enhancement, Chatbot, Topic Summarization, and Topic Suggestions.
Tip: Start with “Fast” and upgrade if needed.
The Fast tier handles most use cases well. You can always switch to a higher tier for specific features where quality matters most, keeping the rest on Fast to conserve credits.
5 Credit Safeguards
Multiple safety mechanisms prevent credit waste, accidental overspending, and billing surprises:
| Safeguard | How It Works |
|---|---|
| Pre-flight credit check | Every operation validates that you have sufficient credits before making the API call. If your balance is too low, the operation is rejected immediately with no charge. |
| Atomic credit reservation | Credits are reserved atomically in a single database operation. Two concurrent requests cannot both pass validation with insufficient balance — preventing overspend race conditions. |
| Automatic refund on failure | If an operation fails after credits are reserved (e.g., network error, model timeout), credits are automatically refunded to your balance. |
| Subscription enforcement | If your subscription expires, remaining credits are locked. They cannot be used until the subscription is renewed — preventing unintended usage. |
| Feature access by plan | Each feature requires a minimum subscription tier. This prevents lower-tier users from accidentally consuming credits on premium features. |
| Low balance warning | When credit usage reaches 90% of your balance, a warning is logged to help you monitor consumption before running out. |
6 Content Optimizations
The system automatically skips content that would waste credits:
| Optimization | What It Does |
|---|---|
| Minimum content length | Posts shorter than 10 characters are skipped during indexing. Empty posts, posts with only shortcodes, or binary data are never processed. |
| AI-generated content detection | Content generated by AI features (e.g., AI reply generator) is flagged and skips AI moderation, avoiding cascading AI processing costs. |
| Already-moderated content | If wpForo’s built-in antispam system already flagged a post, AI moderation is skipped — no need to double-process. |
| Trusted user exemption | Users with a configurable minimum number of posts are exempt from spam detection, reducing moderation credit usage for active community members. |
| HTML & shortcode stripping | Only meaningful text content is sent for embedding. HTML tags, shortcodes, and formatting are stripped first, reducing token usage. |
| Sync state tracking | Already-indexed topics are tracked in the database. Re-indexing operations skip topics that haven’t changed, avoiding duplicate work. |
7 Smart Batching
When indexing large amounts of content, wpForo AI uses batching and scheduling to optimize credit usage:
- Batch processing — Multiple topics are processed in a single API call, reducing per-request overhead.
- Staggered scheduling — Large indexing jobs are split into batches and scheduled via WordPress cron with plan-appropriate delays, preventing rate limit waste.
- Background processing — Indexing runs in the background. You can close the browser and the process continues, with progress tracking when you return.
- Stop button — Active indexing jobs can be stopped at any time from the admin panel, immediately halting further credit consumption.
8 What Costs Credits vs. What’s Free
Costs 1 Credit
- Index a topic (first time)
- Semantic search query (first occurrence)
- AI search enhancement (first query)
- Topic summary (first generation)
- Translation (first language per post)
- Chatbot message
- Image processing (+1 extra per image)
Free (Zero Credits)
- Cached search results (24h)
- Cached translations (permanent)
- Cached topic summaries (24h)
- Cached search enhancements (24h)
- Reading forum content
- Page loads & status checks
- Re-indexing unchanged topics
- Admin panel operations
9 Summary
| Mechanism | Credit Savings | Example |
|---|---|---|
| Feature toggles | 100% per disabled feature | Disable chatbot = zero chat costs |
| Translation cache | 100% on repeat views | Translate a post once, serve from cache forever |
| Search cache (24h) | 100% on repeat queries | “How to install” searched by 50 users = 1 credit |
| Quality tier selection | Variable (cheaper models) | Use “Fast” instead of “Premium” for routine tasks |
| Rate limits | Prevents abuse spikes | Guest limited to N searches/day |
| Content skip rules | 10-20% fewer indexes | Skip empty posts, AI content, already-moderated |
| Batch processing | Fewer API calls | Multiple topics in 1 batch vs. individual calls |
| Credit safeguards | Prevents overspend | Failed operations auto-refund, low balance warnings |
Bottom line: You have full control over which features consume credits. Between feature toggles, multi-layer caching, configurable rate limits, and quality tier selection, wpForo AI is designed to make your credits last as long as possible — while still delivering powerful AI features to your community.