Comparison
Ship your own partner storefronts. Or hire Fermat to build them for you.
Fermat is an agency-style AI commerce service for enterprise brands. Comet is self-serve partner storefronts on Shopify. Both are real; they solve different jobs. Here's how to tell which one fits.
The short version
Fermat is effectively an agency with a SaaS layer. They'll build 3–10 shops a month for you, charge $3,000–$7,000/month, and lock you into an annual contract. Comet is the opposite motion — a Shopify-native product your team, your agency, or your AI agent (Claude Code, Cursor, MCP) operates programmatically, from $49/mo, with unlimited stores and no long-term contract.
If you want done-for-you, Fermat is strong. If you want it done by you or by your agent — with infrastructure that scales across creators, publishers, and campaigns — Comet is the faster call.
Do the math
Fermat Visionary = ~$700 per store per month.
That's $7,000/mo for up to 10 shops. If your creator program has 30 creators, you're either paying out-of-plan or splitting the cohort across quarters. Comet's Growth tier is one-twentieth of that cost and has no cap on how many stores live at once.
Fermat's model makes sense when every shop is a bespoke funnel worth four figures to build. It doesn't scale to programs where the store is a routine artifact you need one of per partner.
Side by side
Plainly. No asterisks.
| Dimension | Fermat | Comet |
|---|---|---|
| Model | Agency-style service — Fermat builds and operates shops for you | Self-serve SaaS — you build and operate stores yourself |
| Who builds your stores | Fermat's team (with your input and brand assets) | You, your agency, your AI agent (Claude Code, Cursor, MCP), or our built-in AI Builder |
| Operator interface | Fermat team via ticketed account management | Dashboard · REST API · MCP for agentic workflows |
| Setup time | 90-day onboarding. Fermat builds your first 5 landers. | First store live in 5 minutes |
| Pricing | Scientist: $3k/mo ($2k for first 120 days). Visionary: $7k/mo ($5k for first 120 days). | $49–$499/mo self-serve · Enterprise tier available |
| Store build limit | Scientist: 3 shops/month. Visionary: 10 shops/month. Extra builds are out-of-plan. | Unlimited stores in every paid tier |
| Contract term | Annual lock-in is standard. 90-day opt-out clause. | Monthly. Cancel anytime. |
| Effective cost per store | Scientist ≈ $1,000/store/month. Visionary ≈ $700/store/month. Rises if you need more than the cap. | Growth ≈ $3–5 per store/month at typical partner-program scale. |
| Platforms | Headless / Salesforce Commerce Cloud / custom stacks | Shopify-native (App Store install) · SFCC · others on request |
| Channels supported | Paid-media funnels (ad-to-checkout), AI search commerce | Creators · Publishers · Campaigns · AI referrals (private beta) |
| Catalogue source | Shop-by-shop curation, often duplicated per shop | Live-synced to your existing Shopify catalogue — one source of truth |
| Dashboard access | Site visits metered: Scientist gets 60/year, Visionary gets 120/year | Unlimited dashboard access for every seat |
| Ownership | Shops live on Fermat's rails | Stores live on your Shopify · your catalogue · your checkout |
| Company | US · enterprise-backed · serves Glossier, BISSELL, ILIA | EU · 200+ partner stores across DTC beauty, fitness, and creator brands |
Fermat details compiled from public sources and third-party reviews. If you see anything we've got wrong, email us and we'll fix it.
Choose Fermat if
You want an agency to build and run it for you.
- You want an agency to build and run 3–10 high-touch funnels a month for you
- Your budget comfortably supports $36k–$84k/year + annual commitment
- You're on a headless, SFCC, or custom stack (not Shopify)
- Your priority is AI-search commerce built and managed as a service
- You prefer outsourcing to hiring in-house, and you want Fermat's team to operate the program
Genuinely — if this is you, Fermat is a good call. Their team is strong and their enterprise customers get real value. Go talk to them.
Choose Comet if
You want to ship partner stores yourself, fast.
- You're a DTC brand on Shopify, $5M–$50M in revenue
- You need more than 10 partner stores a month, not fewer
- You want to move fast — ship a new store in the time it takes to read this page
- You want stores created programmatically — via API, MCP, or your AI agent (Claude Code, Cursor)
- You want infrastructure your team operates directly, not a service you hire
- You want monthly billing with no annual lock-in
The actual moat
Comet is programmatic. Fermat is a team.
Every Comet store is an object that can be created, branded, and launched by an API call or an MCP command. Fermat's stores are created by an agency who schedules a kickoff.
REST API
Your growth engineer integrates Comet into your internal tools. Create a store the moment a creator signs the contract.
MCP server
Your AI agent — Claude Code, Cursor, a custom workflow — talks to Comet natively. "Spin up 20 creator stores from this CSV" becomes a single prompt.
Dashboard
For the partnerships lead who wants to click, not code. Same primitives, same stores — different interface.
An agency's build capacity is capped by their headcount. A programmatic product's capacity is capped by your imagination.
The frame behind it
Agencies and products solve different jobs — and the best answer depends on what your team is hired to do this year.
An agency is right when the problem is one-off, specialised, and worth premium spend: a flagship paid-media campaign, a migration, a headless overhaul. You buy the team.
A product is right when the job is repeatable and you need to scale it across dozens or hundreds of partners without hiring in proportion: a creator program, a publisher roster, a campaign calendar, an AI-agent channel that's growing every quarter. You buy the infrastructure.
Partner stores fall in the second bucket. You're not launching one — you're launching many. That's why Comet is built the way it is.
Ship your first partner store this afternoon.
If Comet fits, you'll have a live store before you'd even book a Fermat intro call.