Get feedback on Cursor projects without a deploy step
Cursor turns a prompt into a working prototype before standup ends. Capsule gets that prototype in front of the people who need to review it — on a live, browser-based link with feedback attached to the page — without you standing up hosting or walking anyone through a screen-share.
The prototype is done; the review hasn’t started
A Cursor-built prototype lives on localhost. To review it, a stakeholder needs a deploy, a screen-share, or a Loom — each one adds friction, and each one detaches the feedback from the artefact. "The button at around 1:40 in the video" is not actionable feedback; it’s a scavenger hunt.
Teams either burn time wiring up staging for throwaway prototypes, or skip review and find the problems after shipping.
How the loop works
Generate
Build the prototype in Cursor — a page, a dashboard, a flow.
Share
Publish to one private browser link — no zip files, no hosting setup.
Review
Reviewers click and comment directly on the live page.
Iterate
Your AI agent reads the feedback and proposes fixes; you decide what ships.
Publish
Approve the version and the same link updates for everyone.
Publish from Cursor over MCP
Capsule installs into Cursor as an MCP server with one command from the install page. Tell the agent to publish, and the generated HTML is live on a private share link in seconds.
Stakeholders review in the browser
They click through the real prototype and leave comments on the element they mean. No accounts to provision for viewers, no environment to maintain, links can be password-protected or time-limited.
Feedback flows back into Cursor
Your agent reads the open comments, proposes a fix per item, and you choose what to apply. The approved iteration is pushed to the same link as a new version — review round two starts where round one happened.
Frequently asked questions
Do reviewers need Cursor or a Capsule account?
No. Reviewers just open the share link in a browser. The link can be public, password-protected or time-limited.
Is this a staging environment?
No — it’s lighter than that. Capsule hosts the generated HTML artefact for review. There’s no infrastructure to maintain: publish, review, iterate, publish. For full-stack apps, teams typically publish the front-end prototype or key flows for design and stakeholder review.
How does the feedback get back into Cursor?
Capsule exposes review comments to your agent over MCP. In Cursor, ask the agent to check feedback: it lists open comments with their context, recommends fixes, and applies the ones you approve.
What formats can I publish?
HTML (self-contained pages, prototypes, dashboards) and Markdown (docs, reports, specs). Each capsule keeps full version history on one stable link.
More review workflows: Review Claude artifacts · Review Lovable apps · Collaborative review for AI-generated apps