A ruttl alternative built around your coding agent.
ruttl lets reviewers comment on a page and even mock up the change they want. Pincushion goes one step further: the change gets made. Each pin is an agent work packet — selector, DOM snippet, screenshot, viewport, thread — that Cursor, Claude Code, or Codex reads via MCP and implements in your actual codebase.
What ruttl does well
ruttl is a design-review tool for websites, web apps, PDFs, and images. Its signature feature is edit mode: reviewers don't just comment, they can nudge elements, swap copy, and change colors on a proxy of the page to show what they mean. That's a genuinely good idea for designer-to-developer communication — a mocked-up suggestion is clearer than a paragraph. It also has video comments, @mentions, and integrations with Jira, Slack, Trello, ClickUp, and Zapier. Pricing is per-user and inexpensive — around $8 per user per month at the time of writing, with a limited free plan (check their pricing page for current numbers).
Where the model differs
ruttl's edit mode produces a picture of the change. Someone still has to open the repo, find the component, and make the real change — the suggestion lives on a proxy copy, not in your code. Pincushion attacks that last mile directly. The pin a reviewer drops on your live app carries the CSS selector, the surrounding DOM snippet, a screenshot, the viewport, and the discussion thread. Your coding agent pulls it via MCP with implement_approved_pins, greps the codebase using the selector, makes the change, and calls fix_and_resolve with the commit SHA, branch, and PR URL. The reviewer's suggestion doesn't become a mockup awaiting a developer — it becomes a diff awaiting review.
Put differently: ruttl compresses the describe step. Pincushion deletes the translate-and-implement step, because the implementer is an agent that consumes structured context, not a person who consumes screenshots.
How Pincushion differs
| Pincushion | ruttl | |
|---|---|---|
| Pin payload designed as agent work packet (selector + DOM + screenshot + viewport + thread) | Yes | Comment + visual suggestion |
| Reviewer can mock up the change visually (edit mode) | No | Yes — signature feature |
| The change gets implemented in your codebase | Yes — by your coding agent via MCP | No — handed off to a developer |
| Center of gravity | IDE (slash commands, branch/PR/deploy linkage) | Web dashboard + PM integrations |
| PDFs / image review | No — live web apps only | Yes |
| Post-deploy verification (AI re-checks the fix) | Yes (Pro) | No |
| Free tier | Unlimited manual pins, 1 project, 25 AI actions/mo | Free plan with tight comment limits |
| Paid pricing model | $19/mo flat Pro (14-day trial); reviewers never pay | Per-user, from roughly $8/user/mo |
ruttl shows developers what the change should look like. Pincushion hands the change to the thing that actually makes it — your coding agent.
When ruttl is the better fit
If your reviewers are designers who want to demonstrate changes visually, and your implementation is done by human developers reading a dashboard, ruttl's edit mode earns its keep — and its per-user pricing is cheap for small design teams. It also covers PDFs and static images, which Pincushion deliberately doesn't. Pincushion wins when the implementing "developer" is Cursor, Claude Code, or Codex: an agent can't read a mocked-up proxy page, but it can act on a selector, a DOM snippet, and a thread delivered over MCP.
From pin to PR, without the handoff
Reviewers are free and unlimited — clients and stakeholders never consume a seat. The free developer tier includes one project, unlimited manual pins, and 25 AI actions a month; every signup starts a 14-day Pro trial with unlimited automation, realtime IDE push, unlimited projects, and access controls. See the loop in Cursor, Claude Code, or Antigravity, read how to review AI-built apps, or compare with Pastel and Feedbucket.