Documentation

Everything you need to install, configure, and use PubSchedule Bridge.

Plugin version 1.2.0

Quick Start

Installation

1. Download pubschedule-bridge.zip from the Download page.

2. In WordPress admin: Plugins → Add New → Upload Plugin.

3. Select the ZIP and click Install Now, then Activate.

4. A new PubSchedule menu appears in your WordPress sidebar.

Initial Configuration

Navigate to PubSchedule → Settings and enter:

  • API Gateway URL — leave as default for cloud deployment.
  • Service Token — copy this from your PubSchedule account dashboard (Dashboard → API Tokens → Create token).
  • Click Save & Test Connection to verify the link.

    LLM Providers

    Adding a Provider Key

    Go to PubSchedule → LLM Providers. Four major providers are supported:

  • Anthropic — Claude models (Haiku, Sonnet, Opus)
  • OpenAI — GPT-4o, GPT-4o-mini
  • OpenRouter — access to 200+ models with one key
  • DeepSeek — affordable high-quality models
  • Paste your API key and click Save Key. The system makes a minimal live test call to confirm the key works before saving.

    Custom Providers

    Use the Custom Provider section for any OpenAI-compatible endpoint — Azure OpenAI, a local Ollama instance, Together AI, or vLLM. You provide:

  • Base URL (e.g. `https://api.together.xyz/v1`)
  • API key
  • Model name (e.g. `meta-llama/Llama-3-70b-chat-hf`)
  • Using the Platform AI Key

    On supported plans, PubSchedule provides a managed AI key so you don't need your own. When generating an article, select Platform AI in the model selector. Monthly usage is tracked on your dashboard and limited by your plan.

    Fallback Chain

    If your primary provider fails (rate limit, outage), the system tries others in order: Anthropic → OpenAI → OpenRouter → DeepSeek. Adding multiple keys gives you resilience against outages.

    Content & Schedules

    Languages

    Tick the languages you want articles generated in. Each ticked language generates one article per scheduled run.

    |------|----------|

    Language selection is limited by your plan (1 language on Starter, 2 on Pro, 3 on Agency).

    Article Source Mode

    After selecting your language(s), you choose the article source mode:

  • News Sources — the system fetches live headlines from your configured RSS feeds and uses them to shape the article. You can select specific sources for this generation, or use all sources for the language.
  • Custom Topic — bypasses the news feed entirely. You write a brief topic description and the AI writes a full article about exactly that topic. No news sources needed.
  • You can switch between modes per-generation and per-schedule — your saved settings are not changed.

    Per-Generation Keyword Selection

    A keyword chip panel appears after choosing your source mode. Each keyword saved for the selected language is shown as a toggle chip:

  • All checked (default) — all saved keywords are included in the article prompt.
  • Partially checked — only the ticked keywords are injected. Useful for keeping the article focused on a specific sub-topic.
  • None checked — treated the same as all checked.
  • When editing a saved schedule, the previously saved keyword selection is restored automatically.

    Schedules

    Create a schedule on PubSchedule → Generation → New Schedule. Each schedule runs automatically at its configured time and generates one article per selected language.

    Schedule options:

  • Frequency — daily, hourly, every N hours, weekdays only, weekends only, specific days of the week.
  • Run time — the clock time (e.g. 06:00) when the schedule fires each day.
  • Source mode — News Sources or Custom Topic (see above).
  • AI model — which provider and model to use.
  • Require Review — when enabled (recommended), articles go to the Article Queue before publishing.
  • Prompt Fragments

    Optional add-ons that extend the generated article:

    |----------|--------|

    Require Review

    When enabled (recommended), all generated articles go to the Article Queue for manual approval before publishing. When disabled, articles are published immediately.

    Independent Article Topics

    What Are Independent Topics?

    Independent Topics are saved briefs that let you generate articles without relying on any news feed. Each topic stores:

  • A title — for your reference only
  • A language — which language the article will be written in
  • A description — the full brief the AI writes from
  • Creating a Topic

    Go to PubSchedule → News Sources and scroll to the Independent Article Topics section. Click Add Topic, fill in the three fields, and save. Topics are stored in your WordPress site.

    Plan Limits

    |------|--------|

    Using a Topic

    When generating an article (Instant or Scheduled), select Custom Topic as the source mode. A dropdown shows your saved topics — select one to pre-fill the description. You can also type a one-off description without saving a topic first.

    News Sources & Keywords

    News Sources

    News sources are RSS/Atom feeds from websites you trust. The aggregator fetches headlines on a configurable interval. It never fetches full article content — only titles and URLs are used as context signals.

    To add a source manually:

    1. Click Add Source and fill in the name, website URL, RSS feed URL, and language.

    2. Set a fetch interval (e.g. every 60 minutes).

    3. Optionally add topic tags to categorise the signals.

    AI Source Discovery:

    Type a topic and click Discover with AI. The AI suggests reputable news websites with their RSS URLs. Click + Add next to any you want to use.

    Source counts are plan-gated: 8 (Starter), 20 (Pro), 40 (Agency).

    SEO Keywords

    Keywords are saved per language and injected into article prompts so the AI includes them naturally. Add them manually or use AI discovery.

    AI Keyword Discovery:

    Type a topic, pick a language, and click Discover. Suggested keywords appear as chips. Add them individually or click Add All.

    Keyword counts are plan-gated: 10 (Starter), 25 (Pro), 50 (Agency) per language.

    Filtering Sources Per Generation

    In the News Sources source mode, a checklist of sources for the selected language appears. Uncheck any source to exclude it from this specific generation. On the next generation, all sources are included again by default.

    Article Queue

    Reviewing Articles

    Every generated article appears in PubSchedule → Article Queue with status Pending Review.

    Each card shows:

  • Title and excerpt
  • Language badge
  • Model used and generation time
  • Source mode (News Sources / Custom Topic)
  • Expand button to read the full article
  • Approving

    Click Approve & Publish. The plugin:

    1. Fetches the full article from the content service.

    2. Creates a WordPress post with wp_insert_post().

    3. Routes the post to the page or category you configured in Settings for that language.

    4. Marks the article as published and stores the WordPress post ID.

    An Edit post ↗ link appears for manual tweaks.

    Rejecting

    Click Reject, enter a reason, and confirm. The article is removed from the queue and marked rejected. Rejection reasons are stored for your records.

    Settings

    API Connection

  • API Gateway URL — the base URL of the backend gateway. Default works with the provided Docker setup.
  • Service Token — your unique authentication token from your PubSchedule account dashboard.
  • Article Posting per Language

    For each output language, configure where the WordPress post should be created:

  • Parent Page — which WordPress page it appears under (optional).
  • Category — which category is assigned.
  • Post Status — `published` (live immediately), `draft` (hidden), or `pending` (awaiting WP editor review).
  • Error Notifications

    Enter email addresses (comma-separated) to receive alerts when any automated step fails: news scraping, LLM call, publishing, keyword check.

    Click Send Test Email to confirm delivery works.

    Log Retention

    Set how many days to keep internal error log entries (1–365 days). Older entries are pruned nightly at 03:00.

    Plan Limits

    Feature Limits by Plan

    |---------|-------|---------|-----|--------|

    Platform AI vs BYOK

  • Platform AI — uses a key managed by PubSchedule. No API key setup required. Monthly quota tracked on your dashboard. Resets at the start of each billing period.
  • BYOK (Bring Your Own Key) — uses your own API key from Anthropic, OpenAI, OpenRouter, or DeepSeek. Daily quota tracked separately. Resets at midnight UTC.
  • You can use both in the same account. Each schedule and each instant generation lets you choose which key source to use.