display.dev

Product Introduction

1. Definition: Display.dev is a gated publishing engine and secure hosting platform specifically designed for AI-generated artifacts. Technically, it is a SaaS (Software-as-a-Service) that provides authenticated static file hosting, a command-line interface (CLI), and MCP (Model Context Protocol) server integration for AI agents.
2. Core Value Proposition: It exists to solve the critical workflow gap of securely sharing sensitive, interactive HTML and Markdown outputs from AI agents (like Claude Code, Cursor, and Codex) within a corporate environment. Its primary value is providing a permanent, authenticated URL with a single command, eliminating the need for complex infrastructure, public exposure, or degraded artifacts like screenshots.

Main Features

1. Gated Publishing Engine: The core technology that hosts arbitrary HTML, CSS, and JavaScript files behind company authentication. It works by accepting file uploads via CLI, web drag-and-drop, or agent-initiated MCP actions, then provisioning a unique subdomain URL (e.g., acme.dsp.so/8f3kx9-q1-review). Access is gated by verifying a viewer's email domain against the publisher's configured company domain or via private artifact settings.
2. Company Authentication (SSO & OTP): Implements secure access control using Google Workspace and Microsoft 365 Single Sign-On (SSO) as well as one-time passwords (OTP). This allows viewers to authenticate with their existing corporate credentials without needing to create a new account, ensuring only authorized personnel can view sensitive agent outputs.
3. Inline Commenting & Agent Feedback Loop: Embeds a collaborative commenting system directly onto published artifacts. Teammates can leave contextual feedback. Crucially, the AI agent that created the artifact can read these comments via MCP, update the source file, republish, and resolve the thread, creating a closed-loop iteration cycle directly within the artifact's URL.
4. Display.dev CLI (dsp): A dedicated command-line tool (dsp publish ./file.html) that enables one-command publishing from any development or AI agent environment. It handles file upload, authentication context, and returns the permanent URL, requiring no git repository, commit, or build pipeline configuration.
5. Unlimited Viewers & Flat Pricing: A key architectural and business model feature. Unlike competitors that charge per seat or per viewer, display.dev pricing is based on storage and features (Pro plan: $49/month). An artifact can be shared with 10 or 10,000 internal viewers without additional cost, removing friction from broad internal sharing.

Problems Solved

1. Pain Point: The degradation and insecurity of sharing interactive AI outputs. Methods like screenshots to Slack lose interactivity, uploading raw HTML files is unusable for non-technical viewers, and public publishing options (like Claude's native button or public GitHub Pages) are insecure for sensitive company data.
2. Target Audience: The primary personas are Developers and Engineers using AI coding agents (Claude Code, Cursor, Codex), Data Scientists/Analysts generating interactive dashboards, and Technical Product/Project Managers who need to review and iterate on specs, prototypes, and reports generated by AI. Secondary users are non-technical stakeholders (executives, legal, design) who need secure, easy access to these artifacts.
3. Use Cases: 1) Sharing an interactive data dashboard or chart (D3.js) built by an AI agent with company leadership. 2) Collaborating on a software design specification or architecture diagram (HTML/CSS) with inline comments from the engineering team. 3) Distributing a sensitive internal performance report or code review summary to a broad internal audience without public exposure.

Unique Advantages

1. Differentiation vs. Traditional Hosting: Compared to Vercel/Cloudflare Pages + SSO, display.dev is 7-40x cheaper for this specific use case, offers a dedicated publish CLI (no git required), and native MCP integration. Versus GitHub Pages, it requires no viewer GitHub accounts. Versus DIY (S3 + Cognito), it eliminates weeks of engineering and maintenance overhead for authentication, commenting, and agent integration.
2. Key Innovation: The deep integration of publishing, authentication, and commenting into the AI agent workflow loop. The combination of a zero-configuration CLI, MCP server for agent actions, and the ability for agents to read and act on feedback directly within the platform transforms a static artifact into a "living document" managed collaboratively between humans and AI.

Frequently Asked Questions (FAQ)

1. How is display.dev different from Claude's built-in "Publish" feature? Claude's publish feature creates a public, unauthenticated URL accessible to anyone with the link. Display.dev creates a gated, private URL that restricts access to authenticated users from your company domain (e.g., @yourcompany.com), making it suitable for sensitive, internal artifacts.
2. Do viewers need to create a display.dev account to see my published artifacts? No. Viewers authenticate using their existing company Google or Microsoft SSO account or a one-time password sent to their email. No separate display.dev account registration is required for viewers.
3. What kinds of files can I publish with the dsp CLI? The primary supported formats are HTML files (with full CSS, JavaScript, and interactivity intact) and Markdown files (which are automatically rendered as styled HTML). It is designed for the static outputs generated by AI coding and documentation agents.
4. How does the commenting and iteration loop work with my AI agent? Teammates leave comments directly on the hosted artifact. If your AI agent (e.g., Claude Desktop via MCP) is configured with the display.dev MCP server, it can fetch these comments, use them as context to update the original source file, republish a new version using the CLI, and mark the comment thread as resolved—all within the same workflow.
5. Why is there a limit on gated artifacts but not public artifacts in the free plan? Gated artifacts require display.dev to maintain and secure authentication infrastructure (SSO, domain verification) per organization, which has a cost. Public artifacts are served without this overhead, allowing for unlimited public hosting on the free tier, which is branded.