You fetch, it reads.
- 1Leave your editor
- 2Find a relevant play
- 3Copy it
- 4Paste it into your agent
- 5Repeat
Loading Motifs…
Connect Motifs once. Your coding agent can search plays from real products, read the evidence and caveats, and use what fits, without you leaving the editor.
Claude Code
CodexCursorVS CodeAny MCP clientAdd a free-trial paywall to onboarding. Check Motifs first.
motifs search_plays query: “trial paywall” funnel_stage: revenue
↳ summaries, strongest evidence first
motifs get_play trial-timeline-paywall
↳ whole record · 7 companies · 9 evidence rows
I’ll use the dated timeline from Trial timeline paywall: today, a reminder before the end, the day billing starts.
COMPANY-REPORTED Foodnoms: the share of trial-eligible new users who made a payment rose 59.6% relative to control in a two-week July 2024 test. Company-reported; the timeline, reminder and other paywall changes were tested together, and cohorts were still maturing.
LIMITS Foodnoms and Built With Science changed other things at the same time, so the effect of the timeline alone is not isolated.
You can still copy any play into a chat. With the MCP, your agent searches Motifs itself, pulls the right research, and keeps going.
Reads over the MCP never count against copies.
A search gives your agent the options. A full play gives it everything needed to judge one properly: the mechanic, real examples, evidence, caveats, adaptations, and sources.
Replace the feature list on the paywall with a dated timeline — access today, a reminder before the end, the day billing starts — then ask to send that reminder, and send it.
The researched sequence, screen details and expected behaviour.
BlinkistMimoFoodnomsOpalBuilt With ScienceSunflowerShareFile
9 entries, each with its findings and limits.
Every limitation and failure mode from the full record.
Guidance for adapting the researched play to your product.
The complete citations, each linked to its source.
A peer-reviewed experiment shouldn't carry the same weight as a company's own case study. Motifs grades every evidence row and records what it does — and doesn't — prove.
Search the library, open a full play, explore a playbook, or apply one to the product you're building.
search_playsFind plays by business type, where they live in a product, the outcome they act on and how strong the evidence is.
search_playbooksFind curated collections of plays for one kind of product.
motifs://catalogEvery value the searches filter on, with how many plays use it.
get_playRead one play whole: mechanism, every case behind it, evidence with its limits, and sources.
get_playbookRead one collection: which plays it holds and why each was chosen.
apply-playbookApply a playbook to what you are building, with your product described by you rather than guessed.
Browse Motifs for free. All-access unlocks full playbooks, complete records through the MCP, unlimited copying, and every new collection we add.
No account, no card
Founding price. $49.99 after the first 100.
Tax may apply
Founding Lifetime closes after 100 members
The library stays free to browse. Full records, playbooks, and MCP access are included with all-access. Founding Lifetime is one payment for all-access, including every curated playbook and the MCP as each ships, and closes after the first 100 members; Lifetime then costs $49.99. Lifetime is the same price everywhere. Checkout by Dodo Payments may show your local currency, and your bank may add a small foreign transaction fee. Before buying, read the subscription terms and refund information.
Paste one command or click one button, then sign in once in the browser. After that, your agent can search Motifs whenever you ask it to.
claude mcp add --transport http --scope user motifs https://motifs.dev/api/mcp/ && claude mcp login motifsNo. Your agent sends Motifs a search or the name of a record, and gets records back. Your repository stays on your machine.
With a free account your agent can still search and compare plays by their summaries. When it asks for a whole record, Motifs tells it that reading is part of all-access.
No. Reading through the MCP never counts against the copies on the site.
Coming soon. The MCP is for looking plays up: your agent searches the library, compares plays by their summaries and reads whole plays and playbooks. The skills, and the Claude Code plugin that bundles them, are parked while we test whether they do more for a build than asking your agent to raise the product decisions a task leaves open.
Claude Code takes one command, and Cursor and VS Code take one click. Codex, the Claude app, ChatGPT and any other client that takes a remote MCP server over HTTP work too: give it the URL and sign in through your browser.
There is no key to copy or leak. Your agent signs in to your Motifs account once, in the browser, so it can read what your plan includes. After that it stays connected.
Consumer-app research on paywalls, trials, onboarding, reminders, goals and stakes, from published experiments and company reports. Each case is graded, and the failures are kept. New collections reach all-access members as they ship.
Browse the library, copy a play into the agent you already use, and see whether it helps what you're building.