Build log

ForgeGEO AI has reached its 100% EarlyForge release.

This release moves ForgeGEO AI from a GEO analysis prototype into a working AI visibility intelligence product for brands, SaaS companies, publishers and website owners that want to understand how AI answer engines interpret, mention, cite and compare them.

The core idea remains simple: traditional SEO shows whether a website can rank on Google. ForgeGEO AI is built for the next layer of discovery: whether AI systems understand the brand, describe it correctly, surface it against competitors and connect it to credible sources.

The 1.0.0 release introduces live multi-engine AI scan support. ForgeGEO AI can now run real server-side visibility scans across OpenAI, Perplexity, Claude and Gemini, store the generated answer, detect whether the target brand is mentioned, identify competitor mentions, extract cited URLs and domains when present, and calculate a visibility score. This turns ForgeGEO from a static diagnosis tool into an active monitoring workflow.

ForgeGEO AI now also includes an AI Engine Monitoring area with live provider execution for OpenAI, Perplexity, Claude and Gemini. Each provider is handled transparently: if an API key, model or provider call fails, the scan records a clear provider status instead of inventing coverage or fake results.

A major part of this release is the WordPress connector foundation. Customers can generate a project-specific connector token, install the ForgeGEO AI WordPress plugin and sync public content metadata from their site into the ForgeGEO dashboard. The plugin sends page URL, post type, title, excerpt, headings, meta title, meta description, schema hints, word count and last modified date using server-side token headers.

This unlocks the new WordPress Content Inventory dashboard. ForgeGEO AI can show how many pages are connected, how much content has been indexed, which schema hints are present, and how fresh the connected content is. The goal is not only to scan AI answers, but to connect those answers back to real pages that can be improved.

ForgeGEO AI now includes a Content Inventory Intelligence layer on top of synced pages. The dashboard evaluates AI Readiness per page, Schema Coverage, Freshness Score, Missing FAQ signals, Entity Clarity and recommended fixes by page. This turns WordPress metadata into an execution queue instead of a passive inventory.

The product can now generate 1,500-word SEO/GEO page drafts in the project language. These drafts include WordPress-ready HTML, FAQ sections, FAQ JSON-LD, Article JSON-LD, metadata, suggested slug, internal link suggestions and publishing notes. A safety guard has also been added to block abusive draft requests such as phishing, fake login pages, credential collection, impersonation, malware or prompt-injection attempts.

ForgeGEO AI also includes premium dashboard outputs: project snapshots, snapshot history, generated reports and llms.txt files. These outputs are stored in dedicated database tables and can be generated from the customer dashboard. This matters because dashboard trends, reports and llms.txt are no longer just visual concepts; they now have persistence and backend logic behind them.

The dashboard has also been upgraded into a more complete GEO intelligence experience. Customers can now see GEO Score, brand mention rate, citation wins, competitor gaps, backlink opportunities, WordPress pages, schema hints, indexed words, fresh content and scan history in one place.

The backlink system has also evolved. Iron customers get curated backlink playbooks based on project market, authority signals and AI citation readiness. Steel and Titanium now add Perplexity-powered backlink discovery, which searches for real web pages, directories, lists, resource pages and niche publications that may be worth reviewing manually. This is not positioned as a full Ahrefs or Semrush-style backlink audit; it is focused on AI visibility execution: finding credible places where a brand could earn authority, mentions or citations that make it easier for AI systems to understand and trust the entity.

This release also introduces a central AI model configuration for ForgeGEO. OpenAI, Perplexity, Claude and Gemini models are managed separately for ForgeGEO, without forcing the same model choices onto other EarlyForge products such as Vendilo or ConvertForge AI. That keeps the architecture cleaner as EarlyForge adds more AI-powered products and lets the team update provider models through environment configuration when models change or become deprecated.

ForgeGEO AI is now marked as 100% on EarlyForge, and the lifetime launch window remains open during the countdown. Customers can still buy lifetime access while the launch window is active. After the countdown ends, the product is expected to move toward annual pricing for new customers.

Since the 100% milestone, ForgeGEO AI has also gained the premium output layer needed for a more serious customer dashboard: project snapshots, KPI history, generated reports, llms.txt outputs, quota enforcement by tier, clean quota notices and SEO draft safety guards. Reports and llms.txt are persisted in the database instead of being static UI placeholders, and the dashboard can now generate real outputs from stored project, scan, content inventory and fix pack data.

The remaining work is now narrower and more operational: test the full workflow on HRYP.com, populate a credible demo project, verify Steel/Titanium backlink discovery end-to-end, add stable Stripe Price IDs for direct lifetime checkout, and continue improving charts, onboarding and scheduled monitoring.

ForgeGEO AI has reached 100%, but the strategic direction is bigger than a completion badge. The product is becoming an operating system for AI visibility: scan what AI engines say, understand why competitors win, improve the pages that matter, and build the citations that make a brand more likely to appear in AI-driven discovery.

ConvertForge AI has reached its 100% customer-ready release inside EarlyForge.

This release moves ConvertForge AI from an early-access conversion assistant into a stable customer workflow: secure widget delivery, OpenAI-powered visitor replies, Website Brain indexing, lead capture, structured lead scoring, hot lead email alerts, customer dashboard, lead inbox, conversation detail pages, CSV exports, WordPress plugin support and project sync tokens for page-update recrawling.

The core positioning has not changed: ConvertForge AI is not just a generic chatbot. The widget is the channel. The product is a conversion intelligence layer for websites that already get visitors but need to understand who is serious, what visitors ask before converting, which objections appear repeatedly and which pages need clearer answers.

The secure widget foundation is now in place. Project keys are restricted to authorized domains, widget requests require trusted browser Origin or Referer context, Redis rate limiting protects public endpoints, OpenAI usage is tracked, and visitor conversations are stored with page URL, referrer, hashed IP, user agent and usage events. This gives site owners a practical way to qualify visitors without exposing admin tokens or private credentials in frontend code.

Website Brain is also active. ConvertForge AI can index authorized public pages, use that content inside widget replies and refresh a single public page when a connected site reports an update. For WordPress sites, the plugin can load the widget and notify ConvertForge AI when posts or pages change, using a secret Project Sync Token that is stored server-side in the plugin settings.

The customer dashboard is now live behind EarlyForge login and license access. Customers can view project metrics, Website Brain status, recent leads, recent conversations, conversation detail pages and CSV exports for leads and conversations. The WordPress plugin ZIP is available from the dashboard, together with project configuration values and clearer warnings about token rotation.

Hot lead email alerts are now working. When a submitted lead reaches hot status, ConvertForge AI can send a Brevo-powered alert to the project owner with lead details, score, urgency, buyer context, suggested next action and follow-up guidance. The 1.0.0 release also improves structured lead scoring so that form submissions with email, company or website context, urgency, budget profile and a concrete business need are scored more accurately.

The final regression pass verified the main production path: ConvertForge API health, EarlyForge health, public widget delivery, HRYP widget injection, widget config, domain enforcement, conversation start, AI message reply, lead capture, hot lead alert event with sent=true, dashboard protection, plugin ZIP download, sync API token protection, WordPress plugin syntax, Website Brain crawl records and absence of admin token exposure in public HTML.

ConvertForge AI is now marked as 100% customer-ready on EarlyForge, but the lifetime launch window remains open until the countdown ends. Customers can still buy lifetime access during this window. After the countdown expires, new customers will move to annual pricing.

The next product step is not another “completion percentage” milestone. It is a controlled demo or short trial flow: limited access for a few days, strict caps on projects, conversations, leads and OpenAI usage, automatic expiry and clear upgrade path to the lifetime launch offer while it is still available, or annual pricing after the launch window closes.

ConvertForge AI has reached 70% completion inside the EarlyForge lab.

This release turns ConvertForge AI from a product concept into a working early-access SaaS foundation: a separate backend, database, OpenAI-powered response engine, embeddable JavaScript widget, conversation tracking, lead capture, multilingual replies, lead scoring, domain enforcement and hardened widget delivery.

The product is being built around a clear business problem: many websites receive traffic, but they do not understand why visitors hesitate, what they are looking for, which questions block conversion and which leads deserve immediate follow-up.

ConvertForge AI is not positioned as a generic website chatbot. The widget is only the channel. The real product is the conversion intelligence layer: qualify visitors, score buying intent, capture structured leads and turn real conversations into practical insights about missing content, unclear positioning, weak calls to action and unanswered buying questions.

At the 70% stage, ConvertForge AI already includes an OpenAI-powered assistant, embeddable widget, project-based managed setup, conversation logging, multilingual response handling, lead capture and lead scoring from 0 to 100. The system can classify leads as cold, warm or hot based on intent signals such as pricing interest, urgency, company context and buying-related messages.

Security work has also started early. Project public keys are restricted to authorized domains and the public API now uses a tighter CORS allowlist. This matters because website widgets can be copied, abused or embedded on unauthorized domains if they are not protected from the beginning.

Customers who enter during this 70% forge window get access to the current early-access build, managed setup for their website and every update up to the completed 100% product at no extra cost. The lifetime price is available for a limited window, and pricing is expected to increase as the product approaches completion.

The next build priorities are customer dashboard, lead inbox, hot lead email summaries, Website Brain crawler, usage limits, self-service onboarding, snippet management, documentation and a clean customer workflow that can support lifetime customers without promising unlimited AI usage.

ConvertForge AI is now public as an EarlyForge product at 70%. The goal from here is simple: harden the product, improve the customer workflow and turn visitor conversations into a reliable conversion intelligence system for website owners, founders, marketers and agencies.

ForgeGEO AI has reached 70% completion inside the EarlyForge lab.

This release turns ForgeGEO AI from a product concept into a usable early-access workflow for AI visibility analysis. The product now has its EarlyForge product page, pricing tiers, Stripe checkout connection, protected app access and the first working analysis flow focused on Generative Engine Optimization.

At this stage, ForgeGEO AI is built around a practical problem: many websites can rank on Google but still remain unclear, uncited or invisible inside AI answer engines. ForgeGEO AI helps site owners understand whether their pages are easy for AI systems to understand, classify and cite.

The current workflow includes project-level analysis, AI visibility checks, structured recommendations and a Perplexity paste analysis feature. Users can paste Perplexity-style AI answers or visibility evidence into the tool and turn that material into a more structured diagnosis.

This 70% stage also adds clearer public FAQs, stronger product positioning and a first blog article explaining why traditional SEO visibility does not automatically translate into AI search visibility.

ForgeGEO AI is not positioned as a replacement for SEO. It is being built as a second layer: a practical GEO tool that helps websites become more understandable, citeable and ready for AI-driven discovery.

The next work is deeper reporting, stronger page-level recommendations, clearer scoring, exportable outputs, better prompt guidance, more onboarding help and final end-to-end testing before the product moves toward 100%.

Vendilo is now at 80% completion. This release adds saved product profiles and a faster output workflow for real launch work.

Customers can now save reusable product context — product name, URL, audience, problem, offer, tone, key features, competitors, forbidden claims, default channels and default language — then use that profile when creating a new campaign. Manual inputs still win, so users can override profile data campaign by campaign.

The campaign detail page now includes copy buttons for the most important launch assets: Google headlines, Meta ads, LinkedIn posts, email copy, short video scripts and landing page suggestions. This keeps Vendilo focused on its intended workflow: generate structured campaign assets, review them, copy or export them, then launch manually with full control.

Direct publishing to Meta, Google, TikTok or LinkedIn remains intentionally outside scope. The next work is interactive checklist persistence, campaign templates, stronger multilingual prompt quality, richer video scripts, admin support tools and final Iron/Steel/Titanium end-to-end testing before 100%.

Vendilo now has a real protected backend for customer campaigns: internal API authentication, account-level tenant isolation, campaign creation, AI generation, Markdown export, HTML export, ads CSV export and monthly usage tracking by successful generation event.

At this stage Vendilo is positioned as an AI Campaign Factory that generates complete launch packs ready to copy, export, review and launch manually. Direct publishing to advertising platforms is intentionally outside the 70% scope.