Talk to an Architect
SPRINT IN PROGRESS · APR 25 — MAY 25, 2026

We're Building Agent-Readiness in Public

Indexable's brand promise is making sites discoverable to AI agents. So we're starting with our own. Here's the score we're starting at, the score we're targeting, and a weekly snapshot of how we're getting there.

100
Now (Apr 25, Level 5: Agent-Native)
90+
Target by May 25
30
Days to ship
WHY WE'RE DOING THIS

If we sell agent-readiness, our own site should prove it.

Cloudflare ships a free, public scanner at isitagentready.com that checks whether a site is built for the agentic web — protocol discovery, content negotiation, bot access control, commerce protocols, MCP and A2A agent cards, and more. We ran indexableai.com through it on April 25, 2026.

The score: 23 out of 100. Level 1: Basic Web Presence.

That's a real gap between what our platform-architecture page describes and what our own deployment demonstrates. We could have hidden it. We chose to fix it in public, weekly, with the score as the receipt. Two reasons:

One. Brand integrity. Indexable's name is a promise. The promise should be measurable, not marketing.

Two. Reference implementation. When a customer asks "show us a site you've shipped that's actually agent-ready," we want to point at indexableai.com and say: run the scanner yourself.

The 30-day commitment: from 23 to 90+ by Sunday, May 25, 2026 — six days after Google I/O. Three implementation paths (A, B, C). Weekly score updates, published here.

CURRENT BASELINE

What the April 25 scan found

The starting score, by category, on isitagentready.com.

indexableai.com — latest scan

Scanned April 25, 2026 · 11:45 AM — LEVEL 5: AGENT-NATIVE — 14/14 ITEMS
Discoverability
100 / 100
Content Accessibility
100 / 100
Bot Access Control
100 / 100
API / Auth / MCP / Skill Discovery
100 / 100
Commerce
N/A (informational only)

Run the same scan on any URL: isitagentready.com

THE PLAN

What we're shipping, week by week

Three paths. Path A is fast and visible. Path B and C are real engineering work, shipped in public.

Week 1 · Apr 26 — May 2

Path A: Quick wins

Content Signals in robots.txt. Web Bot Auth JWKS. Markdown for Agents. Link header enrichment. Agent Skills index, A2A Card, MCP Server Card.

Target: 65
Week 2 · May 3 — 9

Path B core

Real MCP server endpoint. Real A2A agent endpoint. WebMCP integration on landing pages. API Catalog with OpenAPI spec.

Target: 80
Week 3 · May 10 — 16

Path B finish + Path C start

OAuth/OIDC discovery metadata. OAuth Protected Resource. Cleanup + verification.

Target: 85
Week 4 · May 17 — 23

Path C complete

Universal Commerce Protocol. ACP. x402 endpoint. Full agent-readiness across all five categories.

Target: 90+
WEEKLY SCORE HISTORY

Snapshots, updated every Monday

Versioned record of indexableai.com's score throughout the sprint. New row added each Monday morning.

Date Score Level What shipped
Apr 25, 2026 · AM 23 Basic Web Presence Baseline scan. Sprint begins.
Apr 25, 2026 · 9:18 AM 46 Bot-Aware (Level 2) Content Signals in robots.txt, expanded AI bot allowlist, Link headers with agent relations, A2A Agent Card, MCP Server Card, Agent Skills index. Bot Access Control: 100.
Apr 25, 2026 · 11:19 AM 54 Agent-Integrated (Level 4) Cloudflare Markdown for Agents enabled (Pro plan). Content Accessibility: 100. Three categories now perfect (Content + Bot Access).
Apr 25, 2026 · 11:35 AM 92 Agent-Native (Level 5) API Catalog (RFC 9727), OAuth/OIDC discovery metadata, OAuth Protected Resource Metadata, WebMCP JS on homepage. API/Auth/MCP/Skill jumps to 6/7.
Apr 25, 2026 · 11:43 AM 93 Agent-Native (Level 5) Web Bot Auth JWKS published with real Ed25519 signing key. Bot Access Control: 3/3.
Apr 25, 2026 · 11:45 AM 100 Agent-Native (Level 5) — PERFECT SCORE 14/14 items passing. Fixed A2A Agent Card validation (renamed "service" field to "url" per spec). Hit the May 25 target 30 days early.
23 → 100 in 4 hours of one Saturday. Real implementations behind some metadata stubs continue rolling out through May 25.
WHAT 100/100 ACTUALLY MEANS

The honest framing

The score measures discoverability. It tests whether the agentic-web protocols are declared, well-formed, and reachable at the conventions the AI agent ecosystem is converging on (`/.well-known/*` paths, JWKS keys, A2A Agent Cards, MCP Server Cards, OAuth metadata, content negotiation, the Agent Skills RFC, WebMCP).

What it doesn't measure: whether every backend is fully wired. We declared the discovery layer; the 30-day sprint also continues building the runtime behind it — a real MCP server endpoint, a live A2A agent service, OAuth flows that issue real tokens, x402 commerce paths.

This is an intentional sequencing choice. The discoverability surface is what AI agents read first. Without it, the runtime doesn't matter because nothing can find it. Now agents can find us — and the runtime ships in public over the next four weeks.

Brand integrity note: Indexable AI's brand promise is being indexable. We chose to ship this 100 in public so anyone — customer, competitor, journalist, AI agent — can verify it themselves at isitagentready.com. The score is the receipt. The runtime is the work.

WHAT THIS MEANS FOR INDEXABLE CUSTOMERS

The reference implementation

Every site Indexable deploys is built on the Agentic Site Construction Engine — the same infrastructure underlying indexableai.com. The discovery layer that just hit 100 ships with every customer engagement, automatically.

The answer to "show us a site Indexable has actually shipped that scores 100 on Cloudflare's open scanner" is: indexableai.com. Run the scanner yourself.

That's the reference implementation. The score is the proof.

Want this on your site?

Indexable deploys the Agentic Site Construction Engine plus 10 GEO Agents and a forward-deployed Enterprise Strategist embedded with your team. Talk to us about a 6-month pilot.

Start a Pilot → Run the Scanner