Changelog

The current version of AEOForged is v2.2 (August 2026) — Citation Board & Monitoring Ownership. This page records every release in reverse-chronological order, tagged New, Improved, or Fixed.

Release notes and platform updates.

Last updated 8 August 2026

AEOForged current version: v2.2 (Citation Board & Monitoring Ownership, August 2026). Total tools: 30 MCP/API tools. Latest additions: Citation Board, page-tier monitoring ownership, Fix Programme, agent-first self-serve, and Original Research Engine (aeo_study).

Every AEOForged release in reverse-chronological order.
Version	Release	Date	Key features
v2.2	Citation Board & Monitoring Ownership	August 2026	Citation Board, page-tier monitoring accuracy, client + agent self-serve ownership, retainer agent loops
v2.1	Agent-First Packages & Fix Programme	July–August 2026	Fix Programme, agent-first self-serve, ask_human decisions, free-audit & task-pack skills
v2.0	Research Engine, Sector Intel & Honest Measurement	July 2026	Original Research Engine (aeo_study), sector intel, Authority Profile, AI-Ready coverage, local/geo, grounding
v1.9	Launch Hardening	June 2026	Launch hardening: 30 tools, consolidated verbs, async parity, deterministic factcheck, citation honesty
v1.8	Smarter Free Audits + Transparent Audit Depth	June 2026	Smarter free audits (value-ranked coverage) + transparent credit-metered full-site depth
v1.7	Agency Tools + Export Foundation	June 2026	Agency tools + export: client-site cap, org-scoped reads, CSV/PDF/JSON export, in-app AI Visibility
v1.6	Query-First AI Visibility & Citation Monitoring	June 2026	Query-first AI citation tracking, multi-engine monitoring, gap-driven content actions
v1.5	Competitive Edge Suite	May 2026	AI crawlability audit, llms.txt generator, Share of Voice tracker (25 tools total)
v1.2	Composite Workflow Tools	May 2026	Four composite workflow tools (audit-improve, strategy, health, quality-gate)
v1.1	Post-Publish Monitoring & Intelligence Platform	May 2026	Post-publish citation monitoring across 3 engines, alerts, 8-dimension scoring
v1.0	Launch	May 2026	Public launch — 19-tool content intelligence platform via REST, MCP, and web app

August 2026

v2.2: Citation Board & Monitoring Ownership

The client Monitoring tab is now a citation board: questions the client chose, turning green only on a measured page citation. Clients and their agents own the board; room rates, trend, and uptime match that same page-tier honesty. Paid retainers open an agent loop after checkout — setup, monthly strategy delivery, and content lifecycle — without inventing citation guarantees.

New

Citation Board in the client room — branded and goal lanes, empty capacity slots, milestones over filled slots only, soft “cited instead” competitors on non-green rows
Client + agent ownership of tracked questions — room Citation Board APIs and REST prompt create/update/deactivate (history kept; removals free a slot)
Self-serve monitoring controls — Control prompts, optional location bind, goal-lane revenue priority by default, measured competitor accept/dismiss, sector query import for agents
Page vs link estimand — room Monitoring uses page citations for board green, rates, trend, and uptime pips so surfaces cannot disagree
Retainer agent surfaces — post-pay setup, Connect-your-agent for paid Monitor/Grow/Dominate, content lifecycle and content-picks APIs for handoff agents, sector-review confirm for Dominate
Room Articles workspace — list/read/save/score and Ready status for programme drafts; content menu tightened for study-backed secondary picks
Agent trust panel — read / propose / ship_live permissions so GitHub PRs stay review-gated and CMS apply is explicit
Purchases ledger on the client Home tab — stacked package history in one place

Improved

Monitoring accuracy — never-checked prompts prioritized in the check loop; Gemini engine strip counts brand/domain mentions while board green stays page-only
Board change limit — rolling cap on slot composition changes; trend charts label windows that include a revision as directional only
Package entitlement unlocks — Dominate tabs and deliverables follow the active package, not a stale single SKU
Agent readiness probes — deeper MCP / Link / RateLimit checks for live operability reporting
Public site — four-layer homepage and pricing refresh; staff login separated from marketing chrome
Security and reliability hardening across auth redirects, batching, and public surfaces

July–August 2026

v2.1: Agent-First Packages & Fix Programme

Clients connect their own agent with an account API key; Fix Programme delivers a page-first fix kit (developer sheet + agent handoff) without a CMS connection. Free-audit and self-serve skills, human decision gates, and authed task packs make the agent loop discoverable from bootstrap — craft over score on live marketing pages.

New

Fix Programme — page-first work surface after payment: Pages tab, per-page developer sheets, agent handoff self-mint, verify-page loop, optional depth coverage for unpaid deep pages
Agent-first self-serve — dual-door journey (/for-agents and /start), free-audit skill (aeoforge-free-audit), ongoing self-serve skill (aeoforge-selfserve), Fix Programme skill (aeoforge-fix)
aeo_ask_human — durable decision gate (REST /api/v1/decisions + MCP); agents poll until a human answers; wait means stop
Agent task packs — bootstrap next_task routing with sticky active_task; content, authority, visibility, study, crawler-logs, and monthly retainer packs
aeo_feedback — agent feedback channel (optional auth) for product signals; never a remote control channel
World-class Fix Programme handoff contract — page-first claim → apply → verify-page → document_fix with craft-over-score rules on live pages
Score bar on submit-for-review — below-bar submissions rejected unless a human authorizes an explicit low-score acknowledgement
Foundation Sprint delivery — measured page scope and a published score bar before Send; Site module unlock follows the same bar

Improved

Client packages renamed for clarity — Monitor, Grow, Dominate, plus attachable Site module and Fix Programme (storefront and room copy)
Goal-first My packages in the room — eligibility locks, profile confirm readiness floor, and honest wait states for connected work capacity
Free audit path — agents may help edit the brand profile; only a human Confirm plus the trigger phrase starts the audit; hard stop at wait_payment
MCP twins for submit-for-review and gate-status so IDE agents share the same review gate as REST
Milestone-only Updates in the client room — quieter timeline focused on what changed for the client

July 2026

v2.0: Research Engine, Sector Intel & Honest Measurement

AEOForge becomes a measurement and research platform: original studies clients can publish, sector intelligence reports, a seven-category Authority Profile, AI-Ready answer coverage, and local/geo as a first-class dimension. Grounding, branded vs generic citation rates, and Wilson confidence intervals keep every public number honest.

New

aeo_study — Original Research Engine (bench, engine, first-party, and comparison lanes): design → measured run → draft → register own-domain article/methodology/dataset; depth floors refuse thin samples
Sector Intelligence Reports — /intel storefront, measured engine sweeps with confidence intervals, content-war attribution, buyer addendum on tokened copies
Authority Profile — seven-category diagnostic (entity, people, topical, primary evidence, distribution, offsite, AI recognition) with archetype and binding constraint; strategy reports integrate it
AI-Ready Standard + Answer Coverage — capability pillars plus a per-query map of covered / weak / uncovered / offsite (measured citations first)
Local/geo dimension — location-bound tracked queries, per-location scoreboard slices (never blended with global), LocalBusiness scoring, NAP scan via aeo_authority
aeo_citation_sources — which off-site sources real engines cite for tracked queries (Reddit, YouTube, reviews, listicles); opportunities where the client is absent
aeo_crawler_logs — parse an access-log export against the AI bot registry; sample-scoped “never seen” honesty
GA4 AI-referral attribution — join cited pages to AI-referrer sessions (directional correlation, not causation)
Client engagement room — tokenized multi-tab portal (Home, Brand profile, Report, Monitoring, Articles, Socials, Strategy) with Stripe checkout and email rail
Diagnostic funnel (/start) — URL + email → brand profile build → package recommendation → auto-provisioned room
Product dossier on the brand profile — structured own-domain product facts for competitive writing, with required limitations
Monthly strategy reports — Visibility Strategy and Citation Strategy (Dominate) composed from measured inputs; publishable to the client room
Citation Programme delivery — ranked content menu from measured gaps, article skeleton validation, social deployment plan, mark-as-posted with measured asset outcomes

Improved

Grounding gate deepened — research or verified primary sources required at finished boundaries (write, final save, submit-for-review); loud coaching on ordinary saves
Branded vs generic citation split — headline rates never blended; share of voice computed on generic queries when a class map exists
Wilson 95% confidence intervals and noise-floor trend calls on the visibility scoreboard; within-noise moves never alert
Capability scorecard on complete audits — 21 deterministic capabilities; Foundation and Site work quote from capability lines, not ticket counts
Brand profile foundation — durable suppressions, measured site facts and voice mechanics, readiness floor before Confirm
Competitor registry — sector roster ∪ profile ∪ extras, ranked by measured citation evidence
Study design honesty — measured relevance, adversarial critic, bounded claim-only revise loop; methodology is code-authored; negatives are data
Page-type-aware AEO scoring — weights and detectors adapt for docs, commercial, and location pages

June 2026

v1.9: Launch Hardening

30 tools across REST and MCP. Consolidated verbs (create, improve, package, diagnose, factcheck, visibility) compose multiple tools behind a single intent-driven call. Long-running tools (diagnose, visibility, research) run async by default with progress polling. Deterministic factcheck replaces LLM-based verification for statistical claims. Citation honesty tightened: measured page-tier citations kept separate from Brave-search proxy signals.

New

30 total MCP/API tools (was 25)
Consolidated verbs — create, improve, package, diagnose, factcheck, visibility — each composing multiple granular tools behind a single intent-driven call
aeo_factcheck — deterministic grounding check for statistical claims against research sources (free, no LLM, reproducible)
aeo_diagnose — async-default site-wide audit with progress polling via GET /api/v1/tool-runs/{run_id}
aeo_document_fix — capture Remediation Stories (problem + fix) for white-label client deliverables (free)
aeo_import — import external articles into a workspace with baseline scoring (free)
Onboard catalog (GET /api/v1/onboard) now lists all tools with costs, verbs, playbooks, and scoring rubric

Improved

Async parity — diagnose, visibility, and research return 202 + poll_url by default; poll GET /api/v1/tool-runs/{run_id} for progress
Citation honesty — measured page-tier citations kept strictly separate from Brave-search SOV proxy
Grounding gate enforced — research required before finished boundaries; ungrounded saves return coaching warnings
Agent handoff bootstrap includes audit_engagement with queue summary, verify-page instructions, and work access settings
Writing quality measured as AI-tell density (deterministic) — advisory only, never affects AEO score

June 2026

v1.8: Smarter Free Audits + Transparent Audit Depth

The free complete audit now spends its AI analysis on the pages that actually matter — your homepage plus your most valuable commercial and high-authority pages — and the rest get fast technical checks. Going deeper (every page) is a transparent, credit-metered upgrade with the price shown before you run it.

New

Smart free coverage — the free audit deep-analyzes discovered pages in value order (homepage, pricing/commercial, key templates, best-linked) instead of crawl order
Full AI audit toggle — analyse every page (up to 40) as a credit-metered upgrade; the credit estimate is shown before you commit and you're only charged for the pages actually analysed
Per-tier AI Visibility query quotas — track more queries as you scale (each tracked query is re-checked across every AI engine on a schedule)

Improved

Audit reports now state coverage plainly: how many pages got deep AI analysis vs. fast technical checks, with a one-click path to run the full-depth analysis
Honest pricing language — audits are free for the pages that matter; deep full-site analysis is metered rather than advertised as unlimited

June 2026

v1.7: Agency Tools + Export Foundation

Make client sites the agency upgrade lever and ship a real export layer: org-scoped reads so any teammate can pull data, CSV across audits/monitoring/visibility, print-to-PDF, and a one-call per-client JSON archive.

New

Client-site cap — Solo plans manage 1 client site; Agency tiers unlock unlimited (enforced on client + onboarding creation, with an in-app upgrade prompt)
CSV export (?format=csv) for audit scorecards/action items, monitoring status/history/evidence, the query-first visibility scoreboard, and Share of Voice history
GET /api/v1/projects — org/workspace-scoped article list; GET /api/v1/visibility/scoreboard and GET /api/v1/sov as Bearer-auth read APIs
GET /api/v1/clients/[workspaceId]/export — single JSON archive per client (profile + latest snapshot + action items + tracked URLs + visibility), gated to Agency
Print / Save-as-PDF on the agency audit dashboard
In-app AI Visibility page (/visibility) scoped to the active client, with run-check and CSV export

Improved

Owner-only v1 reads (export, project read) are now org-scoped — teammates with an org API key can pull client data
Audit snapshot reads include research context for full-export-capable callers

June 2026

v1.6: Query-First AI Visibility & Citation Monitoring

Reorient citation tracking around the queries that matter — measured citation share and competitors over time, with 'which page won it' as a drill-down — plus multi-engine monitoring and gap-driven content actions.

New

Query-first AI Visibility: curated query sets → measured citation rate + share-of-voice across engines, with winning-page drill-down and suggested queries from sector/research/competitor gaps
Multi-engine citation monitoring (Perplexity, ChatGPT, Gemini) with scheduling, opt-in auto-enrolment from audit snapshots, and a redesigned monitoring dashboard
Query-gap content actions — gap queries become prioritised action items with a content brief
White-label public share links for monitoring reports (client-facing, no AEOForged marks)

Improved

Three-tier citation model (page citation, domain citation, brand mention) with real-evidence status and brand-anchored prompts
Real web grounding for ChatGPT + Gemini answer checks; more resilient provider error handling

May 2026

v1.5: Competitive Edge Suite

Three new competitive-intelligence tools — AI crawlability auditing, llms.txt generation, and Share of Voice tracking — plus rebalanced citability scoring with enhanced extractability weighting.

New

aeo_crawlability — Check AI bot access: robots.txt for 11 AI bots, meta noai directives, X-Robots-Tag, llms.txt, sitemap, JS rendering (free)
aeo_llms_txt — Generate spec-compliant llms.txt + llms-full.txt with semantic page grouping and deployment instructions (3cr)
aeo_share_of_voice — Measure AI answer visibility: SOV %, coverage, avg position, gap keywords across 3 engines for up to 20 keywords vs 5 competitors (10cr)
3 new REST endpoints: /api/v1/crawlability, /api/v1/llms-txt, /api/v1/share-of-voice
3 new MCP tools for competitive intelligence
25 total MCP/API tools (was 22)

Improved

Citability scoring rebalanced — dimension weights adjusted for extractability-focused pages

May 2026

v1.2: Composite Workflow Tools

Four new composite workflow tools that combine multiple pipeline steps into single API/MCP calls with automatic error handling and credit tracking. Available via both REST and MCP.

New

aeo_workflow_audit_improve — Audit a URL and conditionally generate a refresh plan if score is below threshold (5-13cr)
aeo_workflow_strategy — Cluster keywords then research and outline top N pillar topics in one call (~44cr)
aeo_workflow_health — Batch-audit up to 10 URLs with a sorted priority dashboard (5cr/URL)
aeo_workflow_quality_gate — Score content against a target threshold with pass/fail and actionable fixes (free)
4 new REST endpoints under /api/v1/workflow/*
4 new MCP tools for composite workflows
19 total MCP/API tools (was 15)

May 2026

v1.1: Post-Publish Monitoring & Intelligence Platform

Citation tracking across 3 AI answer engines, score decay detection, alert system, monitoring dashboard, and platform proof analytics.

New

aeo_track — MCP tool to register URLs for citation monitoring
Multi-engine citation tracking: Brave, Google AI Overviews, Perplexity
Position tracking with best-position memory across engines
Score decay detection with automatic re-scoring of live content
Alert system — first citation, lost citation, position changes, score decay, content stale
Monitoring dashboard with per-URL detail views and check history
Platform visibility analytics — Brave-search visibility proxy by score bracket (labelled as a proxy, not confirmed AI citations)
Public /proof page — measured answer-engine citation evidence, sample-size and niche-spread gated, reproducible from the evidence ledger
Case study export with user opt-in sharing
Live dashboard updates for long-running tool progress
Feedback loop upgrade — per-section scoring, entity gap detection, competitor-aware suggestions
Workflow UI — guided audit, strategy builder, health dashboard, quality gate
5 new API endpoints for monitoring, alerts, and platform stats
Scheduled citation checks, score-decay alerts, and platform stats aggregation

Improved

AEO scoring now 8 dimensions (added extractability)
Expanded automated test coverage across core scoring and API paths
15 total MCP/API tools (was 14)
Score feedback includes fix priority ranking and specific suggestions
Entity graph intelligence for richer scoring context

May 2026

v1.0: Launch

The first public release of AEOForged — a 19-tool content intelligence platform for AI answer engines, available via REST API, MCP, and web app.

New

19 tools covering the full content lifecycle: create, analyze, maintain, and composite workflows
aeo_research — Live web research with standard and deep modes, geo-targeting, and recency control
aeo_outline — Structured content outlines with snippet targets and word count planning
aeo_write — Section-by-section drafting with verified citations and tone control
aeo_polish — Style pass with brand voice, changelog, and idempotency
aeo_verify — Fact verification with confidence tiers
aeo_score — 7-dimension AEO scoring with extractability bonus (free)
aeo_schema — JSON-LD generation for Article, FAQPage, and HowTo (free)
aeo_extract — Featured snippet, knowledge panel, and PAA formatting with score breakdown (free)
aeo_audit — Fetch any URL, score it, get a prioritized improvement plan with missing entities and schema opportunities
aeo_compare — Head-to-head comparison across all 7 AEO dimensions with margin of victory
aeo_refresh — Fresh research with targeted section-by-section update plan and confidence levels
aeo_cluster — Keyword clustering with pillar/supporting hierarchy, content calendar, and hallucinated keyword detection
aeo_full_pipeline — End-to-end pipeline in draft (12cr) or publish (20cr) mode
REST API with 14 endpoints, structured error codes, and credit-aware billing
MCP server for Cursor, Claude Code, and any MCP-compatible client
Downloadable agent skills (SKILL.md) and pre-built MCP configs
Prompting guide, workflow recipes, and developer documentation
Free tier with 3 projects and unlimited scoring, schema, and extract
Pro tier with unlimited projects and 200 credits/month
Team tier with 5 seats and 1,000 credits/month
Credit packs available for additional usage

Page last updated 8 August 2026