Your AI agent. Our GEO intelligence.
Connect Claude Code, Cursor or Codex to geOracle over MCP. Your agent does the reasoning on the AI subscription you already pay for; we bring the briefings, scoring rubrics, score history, market benchmarks and the review workflow that turn generic AI advice into measurable GEO results.
Free analyze tier. No card required.
$ claude mcp add --transport http georacle https://georacle.ai/api/mcpHow it works
One connection inverts the economics of GEO tooling: the thinking runs on your side, the method and the memory run on ours.
Add geOracle to your agent
One command in Claude Code (or a short config snippet in Cursor and Codex) connects your agent to our MCP server with your team key.
Your agent gets the briefing
For every job, geOracle assembles the exact inputs our own pipeline uses: full product content, market evidence, deterministic signals, score history and the rubric to judge against.
Reasoning runs on your subscription
The heavy thinking happens inside your agent, on your plan. We never resell tokens and never run your content through our models for this work.
We validate, store and remember
Submissions come back through strict server-side validation, land as pending proposals in your dashboard, and build the score history that makes month-over-month progress real.
Why this is the affordable way to do serious GEO
You already pay for a capable AI agent. geOracle charges one flat seat for the orchestration layer: the methodology, the guardrails, the memory and the proof. No marked-up tokens. No per-task fees.
The toolbox your agent gets
Eighteen tools that wrap the same pipeline behind our Shopify app, inverted so your agent does the reasoning. A few highlights:
connect_site
Point geOracle at any site and seed its entity catalog from the sitemap.
get_site_snapshot
Crawl-based site analysis: discoverability, structure and per-page issues.
get_score_briefing
The exact rubric, market evidence and history behind our AI-first GEO score.
submit_improvements
Structure-preserving content edits, validated and queued for review.
get_schema_briefing
Schema.org generation with verified prices, ratings and same-site links.
get_audit_playbook
The full 7-pillar site audit methodology, executed by your agent.
get_geo_news
A curated weekly feed of AI-search changes, turned into concrete tasks.
get_pending_work
The maintenance queue: scheduled refreshes and news-driven work items.
Install the geOracle plugin in Claude Code and /georacle:setup wires all of this, including the weekly routine.
Value you can show your team
The agent works in a terminal. The results show up where your team already looks.
Agent Activity dashboard
Every tool call, submission and applied change, with before and after views your whole team can review and approve.
The premium monthly review
The same monthly email our Shopify clients receive: what was done, how AI visibility moved, and what comes next, written from your real numbers.
History and benchmarks
Score history that stays coherent over time, plus share-of-voice probes across AI engines. Measurement your agent cannot fake on its own.
Who it is for
Builders with Claude Code
Your site's code is already in your editor. Your agent applies improvements directly in the codebase and geOracle verifies them on the next crawl.
Agencies running many clients
One seat per operator, any number of connected client teams. Every client gets their own dashboard, history and monthly report.
geOracle PRO and ENTERPRISE merchants
MCP access is already included in your plan. Mint a key in Team settings and your agent works alongside the app.
Simple per-seat pricing
One seat is one connected person and their agent. Cancel anytime.
MCP Seat
The full pipeline for your agent
Cancel anytime. Access runs to the end of the paid period.
Included in your plan
Every geOracle Professional, Enterprise and Strategic plan includes MCP seats at no extra cost.
Works with your stack
Claude Code is the first-class experience, with a plugin that sets up the whole workflow. Any MCP client with header auth works today.
claude.ai and ChatGPT custom connectors require OAuth; support is on the roadmap. Every terminal agent works today with a key.
Security and data handling
Show-once keys
Keys are hashed at rest, shown once at mint time, revocable instantly, and scoped to your team with per-seat identity.
Your tokens stay yours
We never see, meter or bill your model usage. Briefings go out; validated submissions come back. That is the whole exchange.
Nothing ships without review
Everything your agent submits lands as a pending proposal in your dashboard until a human approves it. Guardrails validate every field server-side.
No training on your data
Your content is used to serve your team, never to train models and never shared across customers.
Frequently asked questions
The Model Context Protocol is the open standard AI agents use to connect to external tools and data. geOracle runs a remote MCP server: add its URL to Claude Code, Cursor or Codex and your agent can drive our entire GEO pipeline with your team's data, history and guardrails.
Claude Code is fully supported, including a plugin with guided workflows and a weekly routine. Cursor and Codex CLI connect with a short config snippet. claude.ai and ChatGPT custom connectors require OAuth and are on the roadmap.
Site connection, crawl snapshots, GEO score briefings and scoring, plus news headlines, with one seat and bounded daily limits. It is a real slice of the product, enough to see your score move. Paid unlocks improvements, schema, audits, automation and the monthly review.
One seat: one team member and their connected agent. Seats are billed monthly through Stripe, prorated when you add seats, and you can cancel anytime with access through the end of the paid period.
No. Professional, Enterprise and Strategic plans include MCP seats at no extra cost. Mint a key in Team settings and connect.
No. The reasoning runs entirely inside your agent on your subscription. geOracle only receives the structured results your agent submits, validates them, and stores them for your review.
A bare agent gives you plausible, generic advice and forgets everything between sessions. geOracle gives it our scoring rubric calibrated against your real competitors, structure-preserving edit rules, verified store facts, score history that keeps results coherent, share-of-voice measurement, and a review workflow so nothing ships unseen. Same agent, very different output.
Every paying MCP team gets the same premium monthly review our Shopify clients receive: what was done that month, how visibility and scores moved, and what to focus on next, grounded entirely in your recorded activity.