The majority of the web's content is now produced by small teams running niche sites, local newsrooms, and specialist publications. For them, content isn't marketing — it's the product.
Yet their workflows haven't changed. A 50-person newsroom process compressed into one person's day.
What got built
AI content market
Jasper, Copy.ai, Writesonic — all targeting marketing teams. Landing pages, ad copy, email campaigns.
Zero tools built for editorial workflows.
What publishers actually do
Built alongside Velora Cycling — a live publication run by us. Every feature was shaped by real editorial needs, not assumptions. We use it every day.
No legacy systems. No “that’s how it’s always been done.” We asked: with a blank slate and today’s technology, what would the best editorial tool look like?
Source trawling, copy-pasting, metadata, internal linking — these are solved problems. So we solved them.
Story selection, editorial strategy, refining the draft — that’s where humans are irreplaceable. Everything else should be a superpower, not a chore.
Six stages, fully automated. The editor steps in twice — to greenlight the story and to review the draft.
Continuous, scored for uniqueness and newsworthiness
Sources, context, and your archive assembled automatically
Written in your voice. SEO metadata and citations built in.
Automated checks before the editor sees it
Lands as a CMS draft. Zero copy-pasting.
Multiple outputs from the same article
SKILLS
Based on the Anthropic standard. Persistent instructions that teach Velora how to write for your publication. Voice, house style, audience, domain knowledge, and editorial rules — applied automatically at every stage.
| Velora | Jasper / Copy.ai | Byword / SEO AI | ChatGPT | |
|---|---|---|---|---|
| Built for | Editorial teams | Marketing teams | SEO farms | Everyone / no one |
| Source monitoring | RSS, email, social | |||
| Research depth | Multi-source, automatic | Basic web search | SEO keyword research | Manual prompting |
| Voice & style system | 6 skill types | Basic brand voice | Manual prompting | |
| CMS integration | WordPress, Sanity | |||
| Validation | Sources, entities, data | |||
| Output quality | Publication-grade | Marketing AI copy | SEO filler | Generic AI slop |
| Human oversight | Two checkpoints, never auto-publishes | User-driven | Optional auto-publish | User-driven |
Former Editor of CyclingNews (the world's largest cycling publication). Former Digital Editor of Rouleur. 15+ years leading editorial teams and content operations.
Former Head of AI at Capital on Tap (London fintech, >50% annual growth for over 8+ years). Built and scaled production AI systems, as well as defining data & analytics strategy.
Friends from cycling. Peter wanted to launch a publication using the best technology available. Danny knew AI could make that possible. One barbecue, one idea, first version live in weeks.
veloracycling.com
6M posts published every day
Mediavine, Raptive, and Ezoic alone
1–10 person teams, no dedicated tech
The opportunity isn’t switching from other AI tools. It’s converting publishers still doing everything manually.
Already validated — paying customers live
Revenue scales with sites — predictable, sticky income
High switching costs — configured sources + trained skills
Natural expansion — publishers add sites, upgrade tiers
Publishers found us on LinkedIn and asked to use the product. No ads. No outreach. No public launch.
Velora Cycling and customer sites are visible, rankable, verifiable. Prospects see output quality before they ever speak to us.
Targeting publisher pain points. We know how to rank content — and we use our own tool to produce it.
Peter's editorial credibility + Danny's AI background. Already our primary inbound channel.
With funding: add targeted paid ads to amplify what's already working organically.
Three hires. 18 months of runway. Clear path to £2M ARR.
Product launch. Hire team. Content engine begins.
Ghost + Webflow CMS. Self-serve onboarding. Case studies live.
Multi-language. Advanced analytics. Enterprise tier. API.