Ladder Flow runs a live Model Context Protocol (MCP) server, so the AI assistant you already use can connect straight to your account and act on your behalf. Research, interview, draft and analyze without copy-pasting between apps.
Works with Claude, ChatGPT, Cursor and any MCP-compatible app. OAuth sign-in, no API keys.
Claude
Ladder Flow MCP · connected
You: i want to record a session about pricing — help me prep Claude used Ladderflow start interview, then Ladderflow get interview angles: Research finished — 4 angles. That ran in the background for about five minutes: two angles from what people are asking online this week, two pulled from your Digital Brain — things you’ve already said but never published. Why founders underprice — you’ve made this point twice in past sessions The “anchor once” rule — your own framework, from the March workshop story Discounting as a trust signal — a contrarian take you flagged in June Each comes with an outline you can edit here. Pick one and I’ll open the live session in your browser. You: what have i already said about discounting? Claude used Ladderflow search brain: Three things in your Digital Brain, in your own words: “A discount you explain is a discount you’ll defend forever.” — session, 12 Aug You think discounting early teaches buyers to wait. You’ve said it three times. A story about losing a deal, then winning it back at full price. All three are marked high-trust in your Brain, so they’re allowed to feed your drafts. None of it is published yet. You: that’s not quite right — i only think discounting hurts early on Claude used Ladderflow update memory: Fixed in your Digital Brain. That belief now reads “discounting early teaches buyers to wait”, narrowed to early-stage deals. Every session and draft from here uses the corrected version. I can also re-file a memory — opinion, framework, story, proof — or change how much it’s trusted. You: write me a linkedin post from the lost-deal story Claude used Ladderflow generate posts: 3 LinkedIn drafts, built from that story and your “anchor once” rule — both straight out of your Digital Brain, in your usual cadence. They’re saved as drafts — nothing is ever published from here. Open Ladder Flow to review and publish, or tell me what to change and I’ll rewrite one now. You: run my friday review Claude used Ladderflow get friday review: 6 posts published this week. Your strongest was “The discount you explain…” — it brought in 4 DMs and a booked call. I can pull the numbers and what came of them any time. Rating the week is still a Ladder Flow job — do it on the Friday Review page and next week’s drafts will lean on the result.
Every tool the server exposes is here, and every one is live today. Pro-only tools are marked.
Search and read the opinions, frameworks, stories, and proof points you've built up — in your own words, not a summary of them. Browse the topics you've covered, or ask for a high-level picture of everything the Brain is holding.
Add a thought by typing it straight into the conversation, or hand it a YouTube, LinkedIn, or X link, a web article, or a PDF to read in the background. Check what has finished, see everything you've added and how much room is left, and delete a source along with everything drawn from it.
A guided three-step flow: research a topic, review the candidate angles Ladder Flow finds, then launch the real voice session — you get a link to open and talk.
List every past voice session, or check the status, phase, and transcript-so-far of any one — including a session that's been launched but not yet started, or one still mid-recording.
Correct a memory that was saved wrong — its wording, how it's filed, or how far it's trusted. Remove one that shouldn't be there at all. And settle the duplicates the Brain deliberately held back instead of deciding for you.
Produce new LinkedIn posts or X threads from a topic, specific Brain memories, or your own custom signal.
Turn an existing high-performing post into something new — same insight, fresh angle.
List, read, and edit your generated LinkedIn posts and X threads directly in the conversation, then get a link to review and publish in-app.
Weekly and range performance snapshots, a read-only view of your Friday Review, and your best-performing patterns and hook formulas. Rating posts and logging pipeline results stay in the app.
Pick your app, copy one thing, and sign in with your Ladder Flow account. No API keys.
claude.ai or the desktop app, on any plan. Free accounts can add one custom connector.
https://mcp.ladderflow.ai/mcp
extension- come from IDE extensions — not your mcp.json. After editing ~/.cursor/mcp.json, reload Cursor (Developer: Reload Window) or use + New MCP Server. ladderflow should then appear under Installed MCP Servers (often as "Needs authentication" until you click Authenticate).Two rules the server follows no matter which assistant is talking to it.
Connecting means signing in with your Ladder Flow account through OAuth 2.1 — a secure, industry-standard login, separate from your LinkedIn and X publishing connections. There are no shared API keys: each connection is scoped to your own account and can be revoked at any time.
MCP can research, draft, edit, and analyze — it never publishes on your behalf. Every post it touches is saved as a draft, and you always make the final call to publish, in-app.
FlowReady when you are
Enjoy up to 120 minutes free. No credit card required.
