Custom pitch
Name a partner. Get a pitch.
Leads for you
Name a partner. Get a pitch.
Name customers. Get more like them.
Other actions
One per line. The studio finds more companies like them and, on a glance, who to call.
A fit score, the one argument, and the roles worth calling. The full pitch — deck, emails, named people — is one press further.
Most useful is context research won’t surface on its own — a live event or deadline, a specific angle to lead with, or a relationship you already have.
The studio prioritizes that person and builds the outreach around them.
A couple of minutes. This keeps running whether or not you watch it.
Your workspace exists the moment you press the button — the research fills in your brief behind you.
Their studio exists the moment you press the button — the research reads their company and fills in the brief behind you. You own the workspace and stay on its access list; invite their team from its settings whenever you are ready.
Added to the workspace's access list by name — just them, never their whole domain — and the workspace is theirs: they own it, you administer it, so you run it for them but cannot delete it or remove them. Leave it empty and the workspace stays yours until you hand it over.
No charge for this during the beta — per-workspace billing arrives later, and you will choose then whether it bills you or your client.
We ask these once.
Who your team already knows — the raw material for warm introductions.
Everything that decides which companies the studio puts in front of you — and every screen that changes one of them writes here.
Who this workspace will never approach, and why. Checked before a pitch starts, before a batch runs, and before anything reaches the Leads shelf.
The inverse pitch: point the research engine at a competitor. What they sell, how they position themselves, who they sell to — and the clients of theirs the run found sources for, each one a prospect you can run next.
Who is the call with, and which call is it? The studio preps you from what it already knows — a pitch from your library rides along whole, a scanned site brings the glance, and a stranger gets a quick look.
Somebody wants a meeting. Paste what they sent — the email, the LinkedIn message — and the studio judges it against what you are actually hunting for: take the call, or it is a waste of your hour. Naming their company is optional; when the studio has pitched or scanned them before, that research weighs in.
What actually happened, per prospect. Move the stage when it moves; dump notes when there is more to say — call notes, a pasted email thread, tidy not required.
Any company — one the studio pitched, or one that walked in on its own. Whatever you write here is what the next run reads.
Your words, verbatim. A loss teaches the next draft as much as a win — a deal that died on implementation time is a thing every later pitch should know.
Who owns each one, how it pays, how much it runs, how far its research got, and whether it is locked. Suspending keeps everything and is reversible; deleting is the same permanent act as the danger zone, and the person to be careful for is the client. This screen administers records — reading a workspace's work still means being on its allowlist.
When each workspace was last opened, everything it has run that spends model budget, how many client workspaces an agency has made, and what all of it has cost us. Quiet means nobody has opened it in 14 days — the ones to nudge. Spend is estimated: our own arithmetic over recorded usage at public list prices, never an invoice. Export as CSV.
MRR ceiling = new MRR per month ÷ monthly churn. Not a health metric — a hard cap on how big this company gets. Only paying workspaces are counted: a comped tester switched off has churned nothing. A cancellation, a failed payment, a chargeback or a deletion is churn; a lock we applied is not. The first 100 days are reported apart from everything after, because workspaces that never got started otherwise drown out the retention of the ones that did. Export as CSV · why they left, as CSV.
The same act as checkout, without the payment: their address gets a workspace, researched while you watch, comped unless you say paid. A company address brings its whole domain; testers come before revenue, by the pricing decision.
Takes the named address off every allowlist carrying it, so it stops resolving to a workspace and can buy or be invited into a fresh one. Domain grants are reported, never touched from here — removing one would lock out everyone at that company, so that is a per-workspace decision on the list above.
Deletes the cached first look for one website — the shared, domain-keyed reading the landing page's URL field writes and every workspace reads. The next person who types that domain gets a fresh scan, so this is how one wrong reading is retired without re-scanning all of them. A malformed reading now re-scans itself on its next hit; use this for one that is merely wrong, or when a sole trader asks for theirs to be deleted. Workspace data — briefs, runs, what a workspace already read — is untouched; removing that is workspace deletion.
What beta testers wrote inside the product — each note carries its workspace, who wrote it, and the screen they were on. Collect for the beta window, then mine the pile for the improvements pass. Export as CSV.
Failed runs shared with “Share this bug with Prospektor” — each carries the plain error the customer saw, the provider’s request id and raw answer, and the run’s last step-log lines.
Every workspace's support chats, whole — what people asked, what the bot answered. Recurring questions are rework signals. Export the questions as CSV.