Menu

Write the method once. Run it in the AI you already use, for every client or brand.

Illustration of a skill in Unilytics: the aeo-audit skill in the Acme Bikes project, with its details, its files (SKILL.md, a reference folder and a data folder) and the first steps of SKILL.md. Beside it, an AI client saves version 13 of the skill and adds a note to the skill's memory.

Unilytics is the skill system for anyone who works through their AI. Keep your methods as versioned skills, with each client's or brand's context beside them, and run both from Claude and Claude Code today.

Connect your AISign in

Your methods and every client's context, in one place

  1. 01

    Every client or brand, its own project

    Each client or brand gets a project with its Brand Kit, knowledge base and collaborators, so every skill runs against the right context.

  2. 02

    Skills as structured files

    A skill is a folder: SKILL.md holds the steps, reference/ the method, data/ the tables and scripts it checks its work against.

    PlannedTurning a chat into reference notes.

  3. 03

    Every save is a version

    Each save becomes a new version with a note on what changed, and a guard stops two edits from overwriting each other.

    PlannedComparing and restoring versions.

  4. 04

    Your AI works the skill's steps

    Your AI finds the right skill, loads the project's context and works through the skill's steps with the research tools.

  5. 05

    The deliverable the skill promises

    The run ends in the skill's output, here a twelve-slide AEO audit deck with speaker notes, ready for the client.

    PlannedSaving a deliverable as a template.

  6. 06

    The skill gets better

    What a run teaches is written to the skill's memory, and the skill reads its memory at the start of the next run.

    PlannedA background agent that proposes the next version for you to approve.

What built-in skills don't do

Claude and ChatGPT can each keep skills now, inside their own app and without your clients' context. Unilytics keeps one library outside any one app, with each client's context beside it.

  • One library, outside any one AI app

    Your skills live in Unilytics, not inside one AI app. Claude and Claude Code load them over one connector today.

    Live

  • Context per client or brand

    Each project's Brand Kit and knowledge base load with the skill, so one method fits every client or brand.

    Live

  • Every version kept

    Each save is a new version, with a check that stops two edits from overwriting each other.

    Live

  • The right skill for the task

    Describe the task and Unilytics returns the best-matching skills across yours, your projects' and ours, each with the reason it matched.

    Live

  • Evals on every skill

    Test cases stored with each skill and run before a new version goes live.

    Planned

  • Skill Lab agents

    Background agents that propose merges, fixes and deeper versions of your skills, for you to approve.

    Planned

For anyone who works through their AI

  • Consultants and agencies

    One method library for every client. Each client's project holds its Brand Kit and knowledge base, so the same skill fits each brand.

    Connect your AI

  • In-house marketing teams

    One brand, many people. The team shares the same versioned skills and brand context, so everyone's AI works the team's way.

    Plans for teams

Setting it up for a team? Developers and the agents they build can read how agents use Unilytics.

Field-tested on our own client work

We built the skill system on our own client work before it was a product, starting with answer engine optimization (AEO). As of 24 October 2026, the founder's library held 58 skills and 943 saved versions, with the keyword research skill at version 85. Eight of those methods ship with Unilytics, scrubbed of client names: four for AEO, the rest for SEO audits, page structures and client documents.

  • Workbook

    AEO keyword research and prompt simulation

    Keyword research for a client, or building the set of prompts to track in AI answers.

    Delivers · A checked workbook with a prompt map and a keyword map, then a baseline and share of voice.

    Live
  • Deck

    Quick AEO audit

    A fast AI-visibility audit or share-of-voice deck for a client or a prospect.

    Delivers · A deck of about a dozen slides with speaker notes, built on a baseline across ChatGPT, Claude and Perplexity.

    Live
  • Prompt matrix

    AEO prompt audit

    Cleaning up a list of tracked prompts that has grown messy.

    Delivers · Keep, rewrite, remove or review on every prompt, with a reason, a coverage matrix and a retire plan.

    Live
  • Deck + workbook

    SEO audit deck and workbook

    An SEO or site audit of a client.

    Delivers · An audit deck, optional workbooks, a 90-day roadmap and a health score.

    Live
  • Clusters

    Prompt-to-page SERP clustering

    Mapping tracked prompts to pages, or finding AEO content gaps.

    Delivers · Prompt clusters assigned to existing or new pages, with an opportunity score per page.

    Live
  • Word doc

    Competitor comparison page template

    "Brand vs rival" and alternative-to pages.

    Delivers · A short Word document with the page order and benchmark screenshots, drawn from 12 published comparison pages.

    Live
  • Word doc

    Industries page structure

    An industries hub, an industry page template, or where industries sit in the menu.

    Delivers · A Word document with wireframes, benchmark screenshots, a CMS field model and build checks.

    Live
  • .docx

    Doc template design

    Client-facing Word deliverables that must look on brand, or use the client's own template.

    Delivers · A branded .docx, with the generator and a render check.

    Live
How we built it on AEO

Every client in context

  • Projects

    Live

    One project per client or brand, so skills always run against the right context.

  • Brand Kit

    Live

    The voice, rules and key facts, returned to your AI in one call with get_project_context.

  • Knowledge base

    Live

    Store the client's or brand's documents, pasted notes and pages with the project.

Illustration of what an agent receives from get_project_context for one client project: the published Brand Kit, with the brand's description, tone of voice and call to action, its content types and its writing rules.

Works with the AI you use

  • One-URL connector

    Live

    Add Unilytics to claude.ai, Claude Desktop or Claude Code with one URL and a Google sign-in.

  • Claude Code sets it up for you

    Live

    Paste the line below into Claude Code. It shows you the command, adds Unilytics after your yes, and you sign in with Google.

  • Claude Desktop with an API key

    Live

    Or connect with an API key from the app.

  • ChatGPT and Grok

    Planned

    The same connector in ChatGPT and Grok.

Paste into Claude Code
Set up Unilytics from https://unilytics.ai/setup.md
Connect your AI

Your AI acts. You stay in charge.

  • Private to you

    Your skills and client context are private to your account and the collaborators you invite to a project.

    Live

  • Nothing erased by one call

    Deleting a skill archives it; nothing you wrote is erased by a single call.

    Live

  • Spending has a ceiling

    Reading skills and client context is free. Research tools use prepaid credits, so spending stops at the balance you put on the account.

    Live

Save the next method you run. Reuse it for every client or brand.

Plans for consultants, agencies and teams are set with you: write to sales@unilytics.ai.

Connect your AIRead this page as Markdown