Lifetime VPS Europe
Lifetime VPS Asia
Lifetime WooCommerce Hosting
Lifetime Hosting
France Lifetime Dedicated Servers
Germany Lifetime dedicated servers
Lifetime Game Dedicated Servers
Chicago, US
Tokyo, Japan
Singapore
Hong Kong
Seoul, South Korea
Amsterdem, Netherlands
London, UK
Stockholm, Sweden
Oslo, Norway
Vienna, Austria
Zurich, Switzerland
Sydney, Australia
Wordpress Hosting
VPS Hosting
DDOS Protection
Submit Ticket
Full Management
Guides
Videos and Podcasts
Discover Creative Ideas
MetaAgentOS High-Ticket Support System
Voxfor Mega Menu
Voxfor Advanced Price Management For WooCommerce
Voxfor Secure Live Chat and AI Support
Voxfor Multilanguage
Voxfor AI Content Summary
WordPress Plugin
Voxfor Autopilot
Autonomous AI SEO system: research, content generation, deep optimization, scheduling, and manual review — bring your own API keys.
Overview
What Voxfor Autopilot is and why you'd use it.
Voxfor Autopilot is a self-contained WordPress plugin that works like an in-house AI SEO team. It handles the full content lifecycle — from keyword research and long-form article generation to intelligent optimization of existing posts — all from your WordPress dashboard, with no external workflow engines, no subscriptions, and no sending your content to third-party platforms.
You supply your own API keys to one or more supported providers; Voxfor handles everything else: prompt engineering, cost tracking, scheduling, change safety, and rollback.
Why Voxfor Autopilot?
Most AI content tools are one-trick: they either write new content or rewrite what you have. Voxfor Autopilot does both, adds a safety net, and puts you in control at every step. The pre-update Review Agent, budget gates, kill switch, and per-post scheduling mean the plugin will never surprise you with an unreviewed change or an unexpected API bill.
AI Content Creation
Full long-form, SEO-ready articles generated in one click with auto internal linking, FAQs, schema hints, and scheduling.
Content Auditor
5-stage Judgment Engine detects content problems and proposes surgical, safe improvements across four editing modes.
Manual Review Queue
Every AI proposal is scored before it reaches you. Approve or reject changes side-by-side with one click.
Keyword Research
Expand any seed keyword into topical variations, questions, and long-tail phrases saved to your library.
Per-Post Scheduling
Configure automation frequency, editing mode, and confidence thresholds per post directly from the post editor.
Budget & Kill Switch
Daily & monthly caps automatically halt all API calls. One-click Kill Switch provides an instant site-wide pause.
Activity Log
Full audit trail of every action, filterable by level (Info, Warning, Error) and searchable by keyword.
Multi-Provider AI
Route each task type (Research, Content, Images, Review) to a different provider for optimal cost and quality.
Requirements
Minimum server and software requirements.
| Requirement | Minimum | Recommended |
|---|---|---|
| WordPress | 6.2 | 6.9 (latest) |
| PHP | 7.4 | 8.2+ |
| PHP Extensions | cURL | cURL, json, mbstring |
| AI Provider Key | At least 1 (OpenAI, Anthropic, Perplexity, or Gemini) | Multiple keys for per-task routing |
| WP Cron | Default WP-Cron | Real server cron |
| SEO Plugin | None required (standalone mode) | Rank Math or Yoast SEO |
Not tested on multisite. Single-site installations are fully supported.
Installation
Two ways to install — automatic (recommended) or manual via FTP.
Automatic Installation
Log in to WordPress Admin
Navigate to Plugins → Add New.
Search for the plugin
Type Voxfor Autopilot in the search box.
Install & Activate
Click Install Now, then Activate. Database tables are created automatically.
Manual Installation
Download the plugin zip
Download
voxfor-autopilot.zip from the plugin repository.Upload to wp-content/plugins
Extract and upload the
voxfor-autopilot/ folder via FTP or Plugins → Add New → Upload Plugin.Activate
Go to Plugins → Installed Plugins and click Activate.
First-Time Setup
Get up and running in under five minutes.
Enter API keys
Go to Voxfor Autopilot → Settings → API Keys and paste at least one provider API key. Keys are stored in your own
wp_options table — never transmitted to Voxfor.Configure Task Routing
Under Task Routing, assign a provider and optional model version to each AI task. Leave "Auto" to use the first configured provider for all tasks.
Set budget caps
Under Budget, set a monthly and daily USD cap. Once a cap is hit, no further API calls are made until the period resets.
You're ready
Use Create Content to generate a new post, or Auditor to optimize an existing one.
You only need one API key to start. Add more providers later to route each task to the best model for the job.
AI-Powered Content Creation
Generate complete, long-form, SEO-ready blog posts in one click.
Navigate to Voxfor Autopilot → Create Content. Fill in the form and click Generate. The plugin runs a research pass first, then uses the Content provider to produce the article.
Creation Options
| Field | Description |
|---|---|
| Focus Keyword | Primary keyword. Used in the title, H2s, and throughout the body. |
| Secondary Keywords | Additional keywords woven naturally into the content. |
| Article Length | Target word count (300–6000). Default 1500. |
| Language | Any natural language. The AI produces all output in this language. |
| Writing Style / Tone | E.g. authoritative, conversational, technical. Free text. |
| Internal Links | Auto-searches your existing posts and inserts contextual internal links. |
| External Reference Links | AI-sourced external references injected as citations. |
| FAQ Section | Appends a structured FAQ targeting long-tail variants of the focus keyword. |
| Author Bio | Appends an author bio block at the end. |
| Include Closing CTA | Appends the CTA template set in Settings → Content Defaults. |
| Affiliate Disclosure | Prepends the disclosure text when monetization type is "affiliate". |
| Images | Optional AI-generated featured image via OpenAI DALL-E or Google Imagen. |
| Video (Veo 3.1) | Optional 8-second B-roll video via Google Veo 3.1. Enable under Settings → Advanced. One video per post. |
| Post Status | draft, pending, or publish. |
| Schedule Date/Time | Optional future publish date. |
The AI never adds a duplicate H1 inside the post body. The plugin automatically strips any H1 the model includes — your post title remains the sole H1 on the page.
Output Format
Generated content is clean Gutenberg-compatible HTML — no shortcodes required. Schema markup hints are embedded in the generated structure.
Content Auditor & Deep Optimization
Audit any existing post and let the AI propose surgical improvements.
Go to Voxfor Autopilot → Auditor, select a post, choose your settings, and click Run Audit.
Problem Signal Detection
- Thin content (low word count or shallow coverage)
- Missing or misplaced focus keyword
- Poor or missing meta title / meta description
- Over-optimized or spammy anchor text
- Missing calls-to-action (CTAs)
- Outdated year or statistic references
- Missing alt text on images
- Missing internal links
- Missing FAQ section
5-Stage Judgment Engine
Observe → Diagnose → Plan → Simulate → Decide
Every proposed change passes through all five deliberation stages before a single character of your post is touched. The AI understands the full context, identifies root causes, plans a minimal set of changes, simulates the post-edit state, and only then commits to an action.
Preserve Map
Before any editing begins, the plugin locks these elements — they will never be modified:
- Gutenberg blocks and shortcodes
- Existing CTAs and internal links
- Factual claims and statistics
- Required phrases specified in the editing context
- Custom HTML you have embedded
Editing Modes
| Mode | Scope | Best For |
|---|---|---|
| Safe Refresh | Meta tags, minor phrasing, keyword density only. | Posts that just need a quick metadata polish. |
| Section Improvement | Targeted paragraph-level rewrites of weak sections. | Posts where most content is fine but a few sections are thin. |
| Guided Rewrite | Rewrites weak sections and resolves all flagged problem signals in one pass. | Posts with multiple issues but a structure worth keeping. |
| Full Rewrite | Full article rewrite; batched to avoid token truncation. | Outdated or poorly structured posts that need a full refresh. |
The Pre-Update Review Agent must score the proposal above your confidence threshold for it to be auto-applied. Proposals that do not pass are held in the Manual Review Queue.
Snapshots & Rollback
Before every optimization a full snapshot of the post content and meta is saved. The rollback endpoint restores it instantly.
Manual Review Queue
Inspect, approve, or reject AI proposals before they touch your live content.
Navigate to Voxfor Autopilot → Reviews. Proposals appear here when the Review Agent score is below the confidence threshold or when the post is configured to require manual review.
| Action | Effect |
|---|---|
| Approve & Apply | Applies the proposed HTML to the live post immediately. Status updates to approved. |
| Reject | Marks the proposal as rejected with an optional rejection reason. |
| Bulk Delete | Select multiple entries and delete them in one operation. |
Keyword Research
Expand any seed keyword into a researched library of topical variations and long-tail phrases.
Navigate to Voxfor Autopilot → Keywords Research. Enter a seed keyword, choose the count, and click Expand.
Keyword Library
- All expanded keywords saved to the internal library (
voxfor_keywordstable). - Each keyword has a status: Pending Used Paused
- Keywords can be clustered into topical groups.
- Track which keywords have already been used to generate a post, preventing duplicate topics.
Keyword Rotation Schedules
Navigate to Voxfor Autopilot → Schedules. Set up automated content creation from your keyword library. The 15-minute cron hook picks up queued keywords and fires content creation jobs automatically.
Per-Post Scheduling
Set optimization automation on a per-post basis from the post editor sidebar.
Every post and page has a Voxfor Autopilot metabox in the editor sidebar. See Per-Post Metabox for the full field reference.
Scheduling Frequencies
| Value | Description |
|---|---|
never | Autopilot disabled for this post (default). |
weekly | Run once per week. |
biweekly | Run every two weeks. |
monthly | Run once per month. |
quarterly | Run once every three months. |
Human-Edit Policy
| Policy | Behaviour when a human edit is detected |
|---|---|
| Block | Skip this optimization run entirely. |
| Warn | Proceed but log a warning to the Activity Log. |
| Ignore | Proceed silently as if no human edit occurred. |
Budget Controls & Kill Switch
Full cost visibility and multiple layers of spend protection.
Budget Caps
- Daily cap (USD) — resets at midnight UTC each day.
- Monthly cap (USD) — resets on the 1st of each month.
- Alert threshold (%) — sends an email when spend reaches this percentage of the monthly cap.
- Alert email — the address that receives budget alert emails.
AI Cost Page
Navigate to Voxfor Autopilot → AI Cost to see a 30-day spend trend chart, per-provider cost breakdown, per-task-type cost breakdown, and per-call log with model name, token counts, and USD cost.
Kill Switch
Instant Site-Wide Halt
The Kill Switch is on the Dashboard and toggled with one click. When active, all outbound AI API calls are blocked site-wide — no jobs start, no optimizations run, no keywords expand. State stored as voxfor_kill_switch.
Activity Log
A full, searchable audit trail of everything the plugin does.
| Level | Meaning |
|---|---|
| Info | Normal operational events — job started, content created, optimization applied. |
| Warning | Non-fatal issues — human-edit detected, low confidence score, retry triggered. |
| Error | Failed API calls, budget exceeded, job dead after max retries. |
- Filter by level and search by keyword.
- Bulk-select and bulk-delete entries.
- Automatic retention: configure in Settings → Advanced (default 90 days).
Debug mode (Settings → Advanced) mirrors all log entries to PHP's
error_log. Disable on production sites.Multi-Provider AI Support
Route each task type to a different provider and model for optimal results and cost.
OpenAI
GPT-5.4, GPT-5, GPT-5-mini, GPT-4.1, DALL-E 3, gpt-image-1/2
Anthropic (Claude)
claude-opus-4-7, claude-opus-4-6, claude-sonnet-4-6, claude-haiku-4-5
Perplexity
sonar, sonar-pro (live-indexed, ideal for research)
Google Gemini
gemini-3.1-pro, gemini-3-flash, gemini-2.5-pro/flash, Imagen, Veo 3.1
Task-Provider Capability Matrix
| Task | OpenAI | Anthropic | Perplexity | Gemini |
|---|---|---|---|---|
| Research | ✓ | ✓ | ✓ Best | ✓ |
| Content Generation | ✓ | ✓ | ✓ | ✓ |
| Image Generation | ✓ DALL-E | — | — | ✓ Imagen |
| Pre-Update Review | ✓ | ✓ | — | ✓ |
| Keyword Expansion | ✓ | ✓ | ✓ Best | ✓ |
| Video (Veo 3.1) | — | — | — | ✓ Veo 3.1 |
The Review task has no fallback by design. If it cannot complete, the proposal is held in the manual queue rather than auto-applied.
SEO Plugin Integration
Native compatibility with Rank Math, Yoast SEO, or standalone mode.
| Adapter | Reads | Writes | How Selected |
|---|---|---|---|
| Rank Math | Focus keyword, meta title, meta description | Focus keyword, meta title, meta description | Auto-detected or forced via Settings → SEO |
| Yoast SEO | Focus keyword, meta title, meta description | Focus keyword, meta title, meta description | Auto-detected or forced via Settings → SEO |
| Standalone | _voxfor_* post meta | _voxfor_* post meta | Active when neither Rank Math nor Yoast is detected |
No hard dependencies on any SEO plugin. If your SEO plugin is deactivated, Voxfor automatically falls back to standalone mode.
SEO Quality Gates
- Reviewer strictness (lenient / balanced / strict) — controls how harshly the Review Agent scores proposals.
- Min score threshold (0–100, default 60) — proposals below this go to manual review.
- Conservative score (0–100, default 85) — above this is considered high-confidence.
- Max articles / optimizations per day — additional rate limits on top of budget caps.
- Disable Scorecard gate — skip the AI scorecard and apply after PHP validation only.
Settings Reference
Complete reference for all settings tabs under Voxfor Autopilot → Settings.
API Keys Tab: api
| Option key | Sub-key | Description |
|---|---|---|
voxfor_api_keys | openai | OpenAI API key. |
voxfor_api_keys | anthropic | Anthropic API key. |
voxfor_api_keys | perplexity | Perplexity API key. |
voxfor_api_keys | gemini | Google Gemini API key. |
Task Routing Tab: routing
| Sub-key | Values | Description |
|---|---|---|
research | openai / anthropic / perplexity / gemini / auto | Provider for research task. |
content | Same options | Provider for content generation. |
images | openai / gemini / disabled | Provider for image generation. |
review | openai / anthropic / gemini | Provider for pre-update review. No fallback. |
keyword_expansion | All 4 providers | Provider for keyword expansion. |
models[task] | Model slug string | Specific model version per task. Blank = provider default. |
fallback[task] | {enabled, provider} | Fallback provider for research, content, images. |
Content Defaults Tab: content
| Sub-key | Default | Description |
|---|---|---|
language | English | Default output language for all AI content. |
article_length | 1500 | Default word count (300–6000). |
writing_style | authoritative | Tone / voice for generated content. |
post_status | draft | Initial post status: draft, pending, or publish. |
cta_template | (empty) | Closing CTA paragraph appended when enabled. |
disclosure | (empty) | Affiliate disclosure prepended when monetization = affiliate. |
Budget & Alerts Tab: budget
| Sub-key | Default | Description |
|---|---|---|
monthly_usd | 50 | Monthly spend cap in USD. |
daily_usd | 5 | Daily spend cap in USD. |
alert_threshold | 80 | Email alert at this % of monthly cap. |
alert_email | (empty) | Email address to receive budget alerts. |
voxfor_kill_switch | 0 | 1 = all AI calls blocked site-wide. |
SEO Settings Tab: seo
| Sub-key | Default | Description |
|---|---|---|
seo_plugin | auto | auto / rankmath / yoast / standalone |
reviewer_strictness | balanced | lenient / balanced / strict |
min_score_threshold | 60 | Proposals below this score go to manual review. |
conservative_score | 85 | High-confidence threshold. |
max_articles_per_day | 5 | Maximum new articles generated per day. |
max_optimizations_per_day | 20 | Maximum optimization jobs run per day. |
retry_count | 1 | API retries before job is marked dead (0–5). |
scorecard_disabled | 0 | 1 = skip AI scorecard, PHP validation only. |
Advanced Tab: advanced
| Sub-key | Default | Description |
|---|---|---|
api_timeout | 60 | HTTP timeout for AI API calls in seconds (10–300). |
batch_size | 5 | Max posts processed per cron tick. |
log_retention_days | 90 | Activity log entries older than this are pruned (min 7). |
debug_mode | 0 | 1 = mirror all log entries to PHP error_log. |
voxfor_enable_video | 0 | 1 = attach an 8-second Veo 3.1 video to generated posts (Gemini only). |
Per-Post Metabox
Configure autopilot behavior per-post directly in the post editor sidebar.
Metabox Fields
| Field | Options | Description |
|---|---|---|
| Enable Autopilot | Checkbox | Master toggle. Must be checked for any scheduled optimization to run. |
| Frequency | never / weekly / biweekly / monthly / quarterly | How often the scheduled optimization runs. |
| Editing Mode | safe_refresh / section_improvement / guided_rewrite / full_rewrite | Depth of changes the optimizer is allowed to make. |
| Task Type | light_refresh / deep_optimization | Light = faster, fewer tokens. Deep = full content audit. |
| Allowed Actions | Checklist (7 actions) | Fine-grain control over exactly which types of changes are permitted. |
| Human-Edit Policy | block / warn / ignore | What to do if a human edited the post since the last run. |
| Confidence Threshold | 0–100 | Minimum Review Agent score for auto-apply. Below this → Manual Review Queue. |
Allowed Actions:
update_meta— Update meta description if missing keywordadd_faq— Add FAQ items (target 5–8)expand_section— Expand thin sections (<100 words)refresh_outdated— Refresh outdated year/stat referencesalt_text— Add missing alt textsrewrite_section— Rewrite weak paragraphsadd_heading— Add new H2 sections
Custom Prompts
Override any AI prompt with your own text. Leave blank to use built-in defaults.
Navigate to Settings → Prompts. Custom prompts are stored in voxfor_prompts and merged over the defaults at runtime.
Available Prompt Keys
| Key | Used For |
|---|---|
research_system / research_user | Research pass before content creation. |
content_system / content_user | Article generation. |
diagnose | Judgment Engine — Diagnose stage. |
plan | Judgment Engine — Plan stage. |
delta | Applying delta edits to existing content. |
reviewer | Scorecard review agent prompt. |
keyword_expansion | Expanding a seed keyword into a list. |
image_prompt | Prompt sent to the image generation provider. |
Supported Placeholders
{language} — Output language (e.g. "English", "German")
{topic} — The article topic / focus keyword
{focus_keyword} — The target SEO focus keyword
{count} — Requested keyword count (expansion task)
{article_length} — Target word count
{writing_style} — Tone / style string
{existing_content} — Full HTML of the current post (optimization tasks)REST API
All admin UI actions are backed by a REST API under the voxfor/v1 namespace.
All endpoints require
manage_options. Base URL: /wp-json/voxfor/v1/Status & Providers
GET/statusPlugin status, version, provider list.
POST/test-providerTest connectivity to a specific provider.
Cost
GET/cost30-day cost data, provider and task breakdowns.
POST/kill-switchToggle the Kill Switch on or off.
Content & Optimization
POST/createTrigger content creation synchronously.
POST/optimizeTrigger post optimization synchronously.
POST/snapshot/rollbackRollback a post to a saved snapshot.
Async Jobs
POST/jobs/startStart an async job; returns a job UUID for polling.
GET/jobs/{uuid}Poll job progress and timeline.
POST/jobs/executeInternal loopback endpoint (token-authenticated).
Keywords & Schedules
POST/keywords/expandExpand a seed keyword into the library.
GET/keyword-schedulesList keyword rotation schedules.
POST/keyword-schedulesCreate or update a keyword schedule.
DELETE/keyword-schedules/{id}Delete a keyword schedule.
POST/keyword-schedules/{id}/run-nowImmediately trigger a keyword schedule job.
Manual Reviews
POST/reviews/{id}/approveApprove and apply a proposal.
DELETE/reviews/{id}Delete a review entry.
Activity Log
GET/activityList activity log entries.
DELETE/activityBulk-delete log entries.
GET/posts/searchSearch posts by title for the auditor post selector.
Database Tables
11 custom tables created on activation. DB schema version: 1.2.2.
| Table | Purpose |
|---|---|
{prefix}voxfor_keywords | Keyword library — seed terms, expanded variants, status, cluster, linked posts. |
{prefix}voxfor_job_runs | Async job records — UUID, type, status, progress timeline, retry count, hash. |
{prefix}voxfor_provider_requests | Per-call API log — provider, model, task type, token counts, USD cost. |
{prefix}voxfor_content_scores | Per-post content scores from the Judgment Engine and Review Agent. |
{prefix}voxfor_post_schedules | Per-post autopilot configuration (frequency, editing mode, allowed actions). |
{prefix}voxfor_keyword_schedules | Keyword rotation schedule configurations. |
{prefix}voxfor_manual_reviews | Pending/approved/rejected AI proposals with proposed HTML and diff. |
{prefix}voxfor_activity_log | Full activity log — level, message, context, timestamp. |
{prefix}voxfor_cost_daily | Daily aggregated spend per provider and task type. |
{prefix}voxfor_snapshots | Pre-optimization post content and meta snapshots for rollback. |
{prefix}voxfor_locks | Distributed lock table to prevent concurrent processing of the same post. |
Uninstalling via Plugins → Delete drops all plugin options and tables. This is irreversible.
FAQ
Common questions and answers.
No. You only need one. You can assign the same provider to all tasks. Adding multiple providers lets you route each task to the best model for that job.
Yes. Standalone mode stores focus keyword, meta title, and meta description under
_voxfor_* post meta. All optimization features work the same way.Only if you configure per-post scheduling and set the editing mode to auto-apply a high-confidence proposal. In manual review mode (the default), every proposed change sits in the queue until you approve it.
Requests automatically retry with exponential backoff. If you have configured fallback providers for Research, Content, or Images, those take over seamlessly. The Review task has no fallback — if it cannot complete, the proposal is held rather than auto-applied.
Yes. The Kill Switch on the Dashboard halts every outbound AI API call site-wide with a single click. Re-enable it the same way.
In your own WordPress database under
wp_options (voxfor_api_keys). They are never transmitted to Voxfor or any third party. No telemetry, no usage pings, no licensing servers.Any language. Set the target language in Settings → Content Defaults. There is no hard-coded language list — if the underlying model supports a language, Voxfor will use it.
No. The plugin automatically strips any H1 block the AI might include. Your post title is always the sole H1 on the page.
Content is generated as clean HTML. Gutenberg blocks are fully preserved. For classic editor or page builders, complex shortcodes are included in the preserve map and will not be modified.
Yes. Set daily and monthly budget caps. Use the AI Cost page to view spend by provider, model, and task type. Use the Kill Switch to halt all activity instantly if needed.
Changelog
Version history and release notes.
1.3.10LatestSecurity
- Fixed: Manual review status now correctly updates to "approved" after Approve & Apply.
- Fixed: Generated post content no longer includes a duplicate H1 tag.
- Hardened: All database queries comply with WordPress Plugin Check security standards.
- Hardened: All settings registered with
sanitize_callback; all$_POST/$_GETinputs sanitized. - Hardened: Nonce verification added to all admin forms and URL-based filters.
- Improved: Activity Log and Manual Review Queue now support bulk-select and bulk-delete.
- Improved: Full Rewrite mode now batches large work plans to avoid AI output token truncation.
- Improved: Guided Rewrite mode now resolves flagged problem signals in the same pass.
- Improved: Content integrity guards re-inject missing Gutenberg blocks and internal links after section rewrites.
- Removed: "Optimize Existing" option from Create Content (use Auditor instead).
- Renamed: Admin menu labels updated to "Voxfor Autopilot", "Keywords Research", and "AI Cost".
1.0.0Initial Release
- Initial release.
Upgrade notice (1.3.10): Important fixes for manual review status and H1 duplication. Security hardening throughout — update recommended for all users.
