EarthLink Network — Products
As we came to run our own operations on AI, one need led to the next—building, monitoring, repairing, and managing—and each of those needs became a product we built in house. What you see here is the suite of products we shaped while practicing team development for the AI era.
Shared components and infrastructure. We carve authentication, notifications, secrets, and billing out of each product and consolidate them in one place.
A secret-distribution platform that securely delivers sensitive files—.env, certificates, kubeconfig, and more—to every device and CI environment with the single command dotvault pull. It provides organizations, permissions, audit logs, and billing, and users do not need an AWS account.
A foundation that consolidates components common to multiple products—block editor, shared DynamoDB logic, rate limiting, notifications, and more—in one place. Each product references it, reducing duplicated development and maintenance costs.
Our own IdP (OAuth 2.0 / OIDC) that consolidates authentication across all products, including promptflow, usesources, alive365, and plovant. It provides everything through APIs—client issuance, token revocation, and multilingual support.
A shared foundation that products can use without implementing their own notification features. It supports prioritization and fallback across delivery channels, with an idempotent API that prevents duplicate sends.
A billing foundation built on Stripe that handles contracts, coupons, and one-time charges through a single API. Payments and cancellations are relayed to each service via signed Webhooks, with automatic retries to ensure reliable delivery on failure.
Site operations, SEO, and news delivery. We manage our outward-facing surfaces together in one place.
A multi-tenant CMS that runs multiple websites on a single foundation and a single admin console. It offers block-level editing, AI-powered page and image generation, and multilingual translation, routing sites by hostname. This page is also delivered by this product.
An operations SaaS that ingests data from Google Search Console, uses RAG to propose SEO improvements, and measures their impact.
A tool that automatically collects the latest information on AI, security, and more, then summarizes, translates, and delivers it in a personalized form. Pre-analysis triage lets it process large volumes of articles at low cost.
Tasks → knowledge → orchestration → AI infrastructure. We run the daily work itself on AI.
A SaaS that lets you work with multiple AIs—OpenAI, Anthropic, Google, and others—across four modes: chat, comparison, debate, and workflow. Outputs are version-controlled, and an iOS version is also available.
An integrated workspace that brings together wiki, tasks, chat, cross-cutting search, and AI agents at the project level. Permissions are tied to each project and follow the addition and removal of members.
An AI development orchestrator that places a local LLM at hand in the decision-maker's seat. It sorts tasks by difficulty, inserts human approval only where needed, and automates everything from implementation and review to merge.
Our own plugin foundation that distributes specification practices, quality standards, and operational discipline to every project. Rules are implemented so that you cannot proceed without meeting them, enforcing verification based on real measurements.
Uptime, infrastructure, and cost. We keep watch so that what we build never goes down—and so that fixed costs never balloon.
Rather than merely checking whether HTTP is alive, it confirms real user operations all the way through in the browser (E2E), monitoring true availability as users actually experience it. It also provides multi-location monitoring, incident management, and status pages as an integrated whole.
An internal infrastructure foundation that reproduces and operates our CI server fleet from code. It bases pass/fail decisions on real-machine kill-tests rather than design documents, ensuring uninterrupted availability even during failures.
A multi-cloud cost management foundation that gives you a single-screen view of AWS, GCP, Azure, and OpenAI costs and notifies Slack / Teams of budget overruns and anomalies. It automates anomaly detection and daily reporting.
Inquiry handling and community. The surfaces that support the people who use what we build.
A foundation that switches FAQs, design, and initial responses per site, letting you operate RAG chatbots for multiple sites on a single foundation.
A community SaaS for supporting AI adoption that consolidates templates, posts, prompt review, courses, notifications, and search into a single foundation. The user-facing and administrator-facing apps are kept separate.
From a single source of truth, we derive tests, manuals, tutorials, and monitoring.
A foundation that generates four things—tutorials, operation manuals, E2E tests, and liveness monitoring—from a single canonical scenario. It maintains tests and documentation consistently from a single definition.
