Sep 2024 to present · Europe
Founder
Visit aranx.com ↗I am building Aranx to put search engine optimization and generative engine optimization on autopilot for founders and small businesses as distribution gets harder, more expensive, and more competitive in the AI era.
Work breakdown
I am building for a distribution reality where paid channels get expensive, classic search is crowded, and AI answers decide which brands get cited before a customer ever reaches a website.
I am putting search engine optimization and generative engine optimization on autopilot: website context, competitor research, market data, topic lanes, keyword selection, briefs, writing, images, internal links, and scheduled publishing in one loop.
I use agentic workflows underneath the product: research agents gather context, planning agents choose winnable opportunities, writing and image-generation agents produce assets, and worker queues turn it into a daily operating cadence.
01
With Aranx, I turn the work of finding opportunities, planning content, generating articles, creating visuals, linking related pages, and publishing on schedule into one repeatable system.
Articles generated
Loading article count
02
I am building the product for a world where visibility means ranking in search results and getting cited by answer engines. Prompt tracking, citations, competitive visibility, and content updates create the feedback loop.
03
Small teams know they need organic distribution, but the work has become a stack of specialist jobs: keyword research, search intent, competitor analysis, internal links, publishing, refreshes, and now generative engine visibility.
04
The technical layer combines Cloudflare Workers, Agents, Durable Objects, Workflows, Queues, Workers AI, AI Gateway, AI Search, AutoRAG, Vectorize, Browser Run, D1, R2, KV, and a search engine optimization data layer that relies on Ahrefs and DataForSEO market signals. I route work across a wide variety of models, balancing speed, cost efficiency, accuracy, agentic capabilities, availability, and open-source availability for each task.
Cloudflare stack
Aranx is built as an agent runtime that happens to publish content, not a page generator with a prompt attached. Cloudflare is the control plane: edge execution, durable agents, browser inspection, retrieval, model routing, retryable jobs, generated assets, and operational memory live close together.
01
Workers and bindings
Workers run the app, webhook handlers, research entrypoints, publishing APIs, and secure bindings to storage, queues, model gateways, browser sessions, and search indexes.
02
Agents and Durable Objects
Cloudflare Agents provide durable identity, sessions, state, scheduling, recoverable execution, and observability. Durable Objects give each website or publishing run one coordinator for locks, memory, progress, and handoffs.
03
Workflows and Queues
Queues absorb bursts and retry small jobs. Workflows model multi-step runs that can pause, retry, resume, and record each step from crawl to research, brief, draft, image, links, review, and publish.
04
Workers AI and AI Gateway
Workers AI handles Cloudflare-native inference. AI Gateway sits in front of model calls for logs, caching, rate limits, retries, fallbacks, spend limits, guardrails, and third-party model routing.
05
AI Search and AutoRAG
AI Search and AutoRAG turn uploaded or crawled material in R2 into an indexed retrieval pipeline with embeddings, search, and grounded answers for website context, older articles, brand facts, and competitor research.
06
Vectorize, D1, R2, and KV
Vectorize stores embeddings for semantic matching, D1 stores relational campaign state, R2 stores screenshots, source documents, images, and generated media, and KV keeps fast configuration signals close to Workers.
07
Browser Run
Browser Run gives Workers and agents controlled browser sessions, screenshots, structured extraction, markdown conversion, crawls, replay, and human intervention when plain fetch is not enough.
08
Model Context Protocol and sandboxed tools
Cloudflare Agents can call Model Context Protocol tools, sandboxed code, browser tools, and AI Search, giving Aranx a path to connect publishing targets, analytics, search data, and review workflows without hard-coding every integration.
09
Observability and guardrails
AI Gateway logs model traffic, token usage, latency, cache behavior, errors, and estimated costs. Agent, Workflow, and Worker observability make it easier to explain why a topic was chosen, which model wrote it, and what failed.