Your customers already work in Claude and ChatGPT. I build the connector that lets them use your product from there: their own sign-in, tools mapped to your real API, an approval step before anything is written, and both directory submissions. Managed for you, end to end, and the code is yours.
Nine of our tools are listed in Anthropic’s Claude directory.
Your product, where they already are
OAuth sign-inRead-only and destructive hintsAn approval step on every writeAudit logEval setAnthropic’s Claude directoryChatGPTThe official MCP RegistryYour repositoryOAuth sign-inRead-only and destructive hintsAn approval step on every writeAudit logEval setAnthropic’s Claude directoryChatGPTThe official MCP RegistryYour repository
Listed in Anthropic’s Claude directory
9
In review with Anthropic
4
In the official MCP Registry
15
Tools Claude can call
69
What your customer types
They ask. Your product does the job.
This is a real exchange with Owner’s Desk, a connector we built and run. Your connector works the same way, with your product’s own jobs: reads answer right away, and anything that changes data waits for a yes.
Who still owes me money? Draft an invoice for the oldest one.
who_owes_meRead only
Reads the open invoices from the owner’s own Stripe account and lists who owes what.
draft_invoiceWrite, held
Drafts the invoice and returns a one-time confirm token. Nothing is sent yet.
confirm_actionApproval step
Runs only after the owner says yes in the chat, and the action lands in the audit log.
How it works / From your docs to the directory
Built on your API, not a guess.
01
I read your docs
Your public API reference and your product pages: which jobs your customers come to you for, and which endpoints already do them.
02
I design the tools
Six to twelve tools mapped to your real endpoints. Each one is marked read-only or destructive, and every write gets an approval step.
03
I build and test it
OAuth sign-in, the tools, an eval set, and every published directory rule checked before anything is submitted.
04
We submit it and keep it working
Docs and privacy pages written, both directory submissions filed, and the code in your repository.
The packages / Set scope, set price, quoted in the conversation
Four ways in. One conversation first.
Every package has a fixed scope and a set price that I put in writing after we talk. Changes to what we built are included, always, with no change order.
Connector Blueprint
For software teams whose own engineers will build the connector.
Your engineers start from a design that already answers what directory reviewers ask.
What you get
The tool design: every tool named and mapped to your real endpoints, each marked read-only or destructive
The auth pattern and the approval step for every write
The directory review checklist for Claude and ChatGPT, worked against your design
An eval set your team runs before every release
A working prototype against your public API
Not included
The production build, hosting and the directory submissions (that is Connector Build)
Changes to your own API
Credited in full toward a Connector Build if you decide we should build it after all.
Your engineers can write this code. What you buy from us is the roadmap your team keeps, and a design that has already been through directory review: the tool names, the read-only and destructive hints, the approval step, the privacy page reviewers read first.
If your team would rather write it, Connector Blueprint hands them that design, an eval set and a working prototype, and it is credited in full if you bring the build to us later.
The proof / Shipped and running
We already build for Claude.
Nine of our tools are listed in Anthropic’s Claude directory. Four more are in review, and fifteen of our servers are published in the official MCP Registry.
Owner’s Desk
Our own OAuth 2.1 connector. It reaches the owner’s Stripe and Google accounts, every write waits for a confirm token, and every action lands in an audit log.
What directory review asks for, how Claude and ChatGPT differ, what customers in your category would do with a connector, and when your own team should build it.
Does a listing mean Anthropic or OpenAI endorses you?
No. A listing means people can find the connector in the directory, nothing more. We build to every published rule and file the submission; Anthropic and OpenAI decide what they list.
Who owns the connector?
You do. The code lives in your repository and runs in accounts you control. We manage it for you end to end if you want that, and your team can run it without us whenever you choose.
What does the approval step do?
Any tool that changes something in your product drafts the change first and waits. Your customer sees what will happen and says yes in the chat before anything is written.
What if a reviewer asks for changes?
We make them. Changes to what we built are included, always, with no change order.
Do you need access to our code?
No. I design from your public API docs and build against your API with the scopes you grant. If a job needs an endpoint you do not have yet, the proposal says so plainly.
One next step
Send me your API docs.
Tell me your product and where your API reference lives. I will come back with the jobs your customers would do in Claude and ChatGPT, the tools I would build for them, and the package that fits.
Claude is a product of Anthropic, and ChatGPT is a product of OpenAI. Modern Mustard Seed is an independent studio and is not affiliated with or endorsed by either.
