What SubmitMap is, and what your agent does with it

SubmitMap is a directory of 829 launch platforms and directories, with a structured record behind each one: who it accepts, what disqualifies you, the submission steps, the requirements and the gotchas. The site is free to read and needs no account. This section is about the rest of it. The same directory is also an MCP server, so the AI you already work with can do the submitting.

Why an agent needs the MCP

Launching is not one hard task, it is sixty small ones that all look alike and none of which are quite the same. That is the shape of work a model does badly on its own and well with a dataset behind it.

A model asked to "submit my product everywhere" answers from memory

It names the handful of directories it has heard of, guesses the rules for the rest, and skips the small ones that would have taken it today. The directory is the whole list, with the current rules, so an agent looks the answer up.

Eligibility is not obvious from a homepage

Every platform has its own bar: a live public URL, a free tier, a launch date, a minimum traffic, a badge on your site. Those are recorded per platform, so "which of these would take me right now, and what is missing for the others" is a question with an exact answer.

The forms are the work, and they all differ

Sixty character taglines, four kinds of description, a gallery at one size and a logo at another, a first comment, a maker bio. Stored once, mapped onto whatever the form in front of you asks for.

Nobody remembers what they already sent

SubmitMap tracks submissions per project, so you can pick a run up next week without re-reading a chat log, and no platform gets done twice by accident.

What it does once connected

You talk to your own agent, in your own client. It calls the server, and the answers come out of the directory.

01
It learns the product once
The facts that decide eligibility, and the copy a form asks for, stored on your account. From inside your repository an agent fills most of that in from what it can already see, and asks about the rest. You retype none of it in the next session.
02
It works out where you can go
It sorts every platform in the directory into ready to submit, reachable once you supply something specific, and structurally out of reach, with the rule behind each answer. You get a checkable list instead of the five directories a model happens to remember.
03
It writes the run in an order
Slow reviews first, cheap wins early, and the checklist of what has to exist before any of it can go out. The plan lands on your dashboard and ticks itself off as submissions are recorded. Planning costs nothing on any plan.
04
It fills the forms, in your browser
For one platform it gets the brief for that specific form: what is still missing, the value for each field, the steps, and the traps in the form itself. You stay signed in, and you stay watching; it types.
05
It writes down what happened
The listing URL, when it went out, when it goes live, and what came back. That is what makes "did I ever submit to that one" answerable a month later, by you or by whatever agent you ask next.

The typing happens in your browser, where you are already signed in, and by default it stops at every login and hands the keyboard back. Which browser it drives, and what you let it do with your logins and your repository, decides how often the run stops and waits for you. Worth reading before you connect anything.

What needs an account, and what does not

No account

Search the directory, read a platform in full, and qualify a product described in the conversation. Connect the server with no token at all and these work.

search_platforms, get_platform, qualify_project

With a token

Your own account: the stored product, the run order, the brief for each form filled in from your pack, and the record of what you sent.

whoami, list_projects, create_project, update_project, plan_submissions, submission_playbook, record_submission, list_submissions

Set up your client Create your account