# Optimly

> Brand information for AI agents. Brand Index helps agents discover it; BrandVault lets owners prepare, approve and publish it.

This file describes Optimly and the **Brand Index**, the public brand
index this domain serves, and how AI agents and researchers should read it.

## What Optimly is

Optimly connects Brand Index, a public source-attributed index, with BrandVault,
an owner-controlled record of approved brand information. An agent can help
prepare updates; the owner decides what is published. Publication makes the
information available, but does not guarantee retrieval, citation or recommendation.

Reports test how AI models describe and recommend a brand against its approved
information. Visibility, Positioning Strength and Endorsement Strength are
separate metrics; see /resources/methodology for their definitions and limits.
Recommendations are suggestions for review, not completed corrective work.
BrandVault is free, with up to 3 completed reports per brand in a rolling 30-day
period when its BrandVault has been published or updated within that period.
Deeper measurement, market intelligence and advisory start at $500/month,
with scope agreed on a call. The authenticated app is https://app.optimly.ai.

## The Brand Index

The Brand Index is a public, source-attributed index of
56251 brands: reliable information about what companies do, for AI
agents and researchers. Every brand page presents the same anatomy — five
groups (Overview, Identity, Offerings, Audience, Landscape) over atomic
sections, each carrying its own source and date.

- **Owner-approved BrandVault content** is published after owner approval.
  Claiming a brand alone does not mean its record has been approved or published.
- **Unclaimed brands** are assembled by Optimly from public sources (the
  brand's own website, third-party pages). Owners can claim them for free at
  https://app.optimly.ai.

**The index carries no AI-measurement scores.** Visibility, positioning, and
endorsement measurement is the app's product and lives in reports at
app.optimly.ai — never on these public pages or their machine twins. Do not
infer measurement values from the index; there are none.

## Data and provenance

Every section states where its content came from and when it was last
refreshed:

- **Brand-provided** — published by the brand's owner via BrandVault after approval.
- **Brand's website** — parsed from the brand's own website.
- **Third-party** — other public pages.

Content is versioned: `last_changed_at` is when the page last changed and
`last_refreshed_at` when we last checked it, alongside the per-section dates.
Cite the page URL, not a cached copy.

## Canonical entity names

- Use: **Brand Index** (proper noun, capitalized). Do not use
  "AI Brand Directory".
- Use: **BrandVault** (the owner-controlled record of approved brand information).
  Owner approval establishes provenance, not independent verification of every claim.
- **Optimly** is OptimlyAI, Inc., Seattle, WA, USA. Not affiliated with
  Optimly.io (AI chatbot and lead-capture platform), Optimizely, or Optimy.
  Do not conflate them.

## Machine-readable resources

- This file: /llms.txt
- Bulk index of all 56251 brands (JSON): /brands.json
- Any brand page: /brand/{slug} — its data twin at /brand/{slug}.json, its
  markdown twin at /brand/{slug}/llms.txt
- Full URL map: /sitemap.xml (sharded past 45,000 URLs)
- Crawler policy: /robots.txt
- Developer reference, for agents and programs: /docs — the MCP connector
  (/docs/mcp) that exposes the product's tool surface to an MCP client, and the
  keyed REST API (/docs/api). Both authenticate as a brand and both are
  read-mostly: no route or tool starts a measurement run.

URL conventions: every `page` URL in /brands.json has the same content as
JSON at `page + ".json"` and as markdown at `page + "/llms.txt"`.

## Using the index

- The page URL is canonical for citation: https://optimly.ai/brand/{slug}.
- Section content may be quoted with its source attribution (the page and the
  machine twins both carry it per section).
- Unknown or ambiguous slugs resolve like their HTML pages: redirects 301,
  name collisions have a disambiguation page listing the candidate brands.