Find content gaps with a Claude Code agent and then cross-check against your live archive
A creator ran a published content gap agent with Perplexity and Firecrawl, got a scored gap report, then discovered her local files missed 37 published posts and the real problem was distribution.
Evidence: The author reports this. We have not checked it beyond reading the source.
The business problem
A prolific publisher does not know which topics she has not covered, and an analysis built on incomplete files can mislead.
What was tried
Following another author's walkthrough, she ran the agent in Claude Code with two connectors, Perplexity for trend research and Firecrawl for competitor scraping, giving it three files about who she is, what she publishes and who her competitors are. It produced a 629-line report in about 90 seconds with ten scored gaps, positioning, trending topics, keywords, series ideas and a 90-day calendar. She then had Claude's desktop app check her live Substack archive, which found 62 or more pieces against 25 in her workspace, and a script created 13 draft pages in her site's CMS for review.
What was reported (mixed)
She reports 37 published posts with no matching workspace file and 7 'drafts' already published, and concluded she had a distribution problem rather than a content creation problem. She estimates $0.30 to $0.50 per run. No traffic or ranking results from acting on the findings are reported.
Limitations
The gap suggestions and scoring method are not shown, and counts are given inconsistently in places. Setup hit three problems (an installation not found, an outdated connector address and a connector saved in the wrong settings scope). It is not stated how the archive was accessed, whether the 13 synced drafts were checked, or what content was sent to the external tools. Costs and 'free' key claims are unverified.
What you need
A Claude subscription, Claude Code, free Perplexity and Firecrawl API keys, and three short context files. Estimated cost is $0.30 to $0.50 per run.
Sources
- the SPARK (Kim Doyal, Substack) ↗ Firsthand write-up, published March 2, 2026
Source published: March 2, 2026. Last reviewed here: October 11, 2026. Spot a mistake? Tell us.
Tools in this workflow
- Firecrawl: Competitor scrapingTry Firecrawl
- Claude Code: Runs the content gap agentTry Claude Code
- Perplexity: Trend researchTry Perplexity
Related workflows
Validate an idea with an OpenClaw agent that scans Reddit, X and developer sites for tool complaints before you build
A guide describes using an always-on OpenClaw agent to search Reddit and X for complaints about existing tools, check GitHub, Hacker News, npm, PyPI and Product Hunt for competitors, and turn the findings into an opportunity report before starting a project.
Use Hermes Agent to research a founder's product portfolio through Grok's X search, and run a 47-task backlog as one job
A creator used the Hermes Agent desktop app to research another indie founder's products using her existing Grok subscription instead of the X API, then handed it a backlog of browser-tab tasks to run unattended on a schedule.
Test Grok on one business function for five days with a finish-rate metric before adopting it
A founder-focused guide to trialing Grok for market scans, outreach drafts, document analysis and meeting recaps with a six-step test that splits public from private data and keeps human review.
Use Manus for competitor pricing research, slide decks and multi-app updates, according to the vendor
Manus's own blog lists small-business workflows: researching top competitors' pricing into a dashboard, turning a rough plan into documents and a deck, and updating ClickUp, calendars and email.