Beyond One-Off Extractions
Spectre Extract is powerful for on-demand intelligence. But most real-world use cases are not one-off — they are ongoing. You do not want to check a competitor's pricing page once. You want to know when it changes. You do not want to research a market once. You want a weekly briefing.
Spectre Campaigns are how you turn recurring intelligence work into an automated operation.
What a Campaign Is
A Campaign is a named, configured intelligence operation that runs on a schedule against one or more target domains or URLs. You define it once; the platform runs it continuously.
Each campaign has:
- A target list — one or more URLs or domains to crawl on each run
- A crawl configuration — depth, page budget, and output format per target
- A schedule — cron expression or human-readable interval (daily, weekly, every Monday at 08:00)
- An output handler — where results go after each run (Nexus Report, cascade trigger, or both)
- A diff mode — optional change detection that compares the current crawl to the previous one and highlights what changed
Change Detection
The most powerful feature of Campaigns is change detection. When diff mode is enabled, each run produces a structured comparison:
- Added content — text that appears in the current crawl but not the previous one
- Removed content — text that was in the previous crawl but is gone
- Modified headings — structural changes to the page's section hierarchy
- New links — external or internal links that appeared since the last run
- Removed links — links that have been removed
This means your campaign does not just tell you what a page says — it tells you what changed since last Tuesday. For competitive monitoring, this is the difference between having a live intelligence feed and having a static snapshot.
Cascade Integration
Campaigns are designed to feed into Nexus Cascades automatically. When a campaign run completes, it can trigger a cascade that processes the results.
A typical pattern:
- Campaign runs on schedule, crawls three competitor domains
- Results land in PocketBase as structured records
- Event cascade fires — watching the campaign_results collection
- Cascade step: agent — the Librarian synthesises the results into a competitive summary
- Cascade step: db_write — the summary is written to a Nexus Report
- Cascade step: http — a notification is sent to your Discord channel
The entire chain is automated. Monday morning, you have a competitive intelligence briefing in your notifications. You did not do any of it manually.
Managing Campaigns
Campaigns are managed from the Spectre section of the dashboard:
- Create — name your campaign, add target URLs, set the schedule, configure crawl settings
- Run manually — trigger a one-off run outside the schedule to validate your configuration
- History — every run is logged with its timestamp, target count, pages extracted, and any errors
- Pause/resume — temporarily disable a campaign without deleting its configuration or history
- Archive — mark a campaign as complete and remove it from the active list while preserving its data
Use Cases
Weekly competitor monitoring — crawl five competitor product pages every Monday, run a diff to see what changed, file a Nexus Report with a summary.
Pricing intelligence — track pricing pages across your competitive set, alert on any change detected.
Content gap analysis — crawl industry publication archives weekly, identify topics competitors are covering that you are not.
Regulatory monitoring — crawl regulatory body websites daily, alert on any new publications in sections relevant to your industry.
Partner intelligence — monitor partner and client websites for signals of strategic changes — new product pages, new leadership, new market focus.
Job listing intelligence — crawl competitor careers pages weekly to understand what they are building by what they are hiring for.
The Campaign vs. the Cascade
Campaigns and Cascades are complementary, not competing.
A Campaign handles the collection layer — scheduled, multi-target web crawling with change detection. It is the data acquisition pipeline.
A Cascade handles the processing layer — taking the collected data and running AI analysis, report generation, notification, and action on it. It is the decision and output pipeline.
The Sovereignty Protocol is designed so these two systems connect naturally. The output of a Campaign is the input of a Cascade. Together they form a complete intelligence-to-action pipeline that runs without manual intervention.