Display.dev Raises EUR 470K to Give AI Agents a Shared Document Layer
Ott Ilves and Carl Rannaberg are building a company-gated URL system for the HTML and Markdown files agents already produce.
By Ryan Merket · Published
Why it matters
Agent products are generating more work than teams can review cleanly, and Display.dev is betting the collaboration layer around those outputs can become its own software category.
Display.dev, the Tallinn-based agent document platform founded by Ott Ilves and Carl Rannaberg in April 2026, has raised EUR 470,000 in pre-seed funding, Tech.eu reported on July 7th.
The round was backed by Outlast Fund, FIRSTPICK, Curiosity VC, and angel investor Henrik Bohman, whom Tech.eu identifies as Wise's first product manager. Display.dev has not disclosed a lead investor, valuation, customer count, revenue, or usage metrics.
Ilves and Rannaberg are entering the agent tooling market from a narrow wedge: agents can now produce reports, specs, dashboards, slide-like HTML pages, and Markdown documents, but teams still review much of that output through screenshots, file attachments, Slack threads, or whatever interface generated it. Display.dev's bet is that the missing surface is less glamorous than the agent itself. It is a durable URL, company authentication, comments, version history, and a way for the agent to read feedback and publish the next version.
The wedge is the artifact, not the editor
Display.dev is not trying to replace Google Docs for humans writing from scratch. Display.dev is built for artifacts already made by Claude Code, Codex, Cursor, CI systems, or other agents that can write HTML or Markdown files. The Display.dev homepage describes the product as "A URL for everything your agents make" and says teams can publish from CLI, MCP, or API, then put the result behind Google or Microsoft login, with one-time codes as another access path.
That design choice matters because most AI-agent output does not behave like a traditional document. A report may include live charts. A planning artifact may be HTML. A product spec may arrive as Markdown. A benchmark or internal analysis may be a rendered page that loses context when pasted into a chat window. Display.dev's pitch is that those files should remain rendered as built, stay attached to comments, and keep their history after a team switches from one agent or model vendor to another.
The company's docs show four publishing paths: terminal, AI assistant, browser upload, and CI pipeline or app. The same docs say Display.dev turns any HTML or Markdown file into a stable URL gated by company auth. In practice, that makes Display.dev closer to a review and sharing layer for machine-generated work than a general document editor.
Pricing attacks the seat-license trap
Display.dev's pricing is also part of the product thesis. The homepage says every plan includes unlimited viewers, with pricing based on storage and capabilities rather than per-seat access. The free tier includes 10 gated artifacts, unlimited public artifacts with Display.dev branding, and 50MB of storage. Solo costs USD 15 per month and adds unlimited gated artifacts, 1GB of storage, and 10 versions. Pro costs USD 49 per month and adds Google and Microsoft SSO, private artifacts, 25GB of storage, 50 versions, and 90 days of audit logs. Enterprise starts at USD 499 per month with custom storage, unlimited versions, and 365 days of audit logs.
That is an explicit challenge to a familiar enterprise software failure mode: a tool becomes expensive once everyone who needs to read, comment, or approve something needs a seat. Display.dev is trying to sell the opposite motion. Agents generate the artifact, a small number of builders publish it, and everyone else at the organization can view and comment without turning into a licensed user.
Display.dev's own competitive table compares the product with Vercel Pro plus SSO, Cloudflare Pages plus Access, GitBook Ultimate, GitHub Pages private, and a do-it-yourself S3 plus Cognito setup for the job of sharing one HTML artifact with 100 viewers behind company authentication. Display.dev claims it is cheaper for that task, though that comparison is Display.dev's positioning rather than neutral market data. The useful signal is the list of alternatives. Rannaberg and Ilves are not pitching against word processors alone; they are pitching against the hacked-together stack that teams use when agents create files that need to be private, reviewable, and persistent.
Early proof is narrow, and that is the point
Display.dev has one named customer story on its site: Indigo Engineering. In the case study, CEO Max Wood says that before using the product, the team shared artifact files in Slack and commented through screenshots, especially for HTML artifacts. The case study says Indigo now uses Display.dev to share and collaborate on Markdown and HTML artifacts.
That is a small proof point, not evidence of broad market adoption. It still shows the specific workflow Display.dev is trying to own: agent-generated output gets created quickly, then human review slows down because the output has no obvious home. The company's opening round gives Ilves and Rannaberg enough capital to keep building around that workflow before the larger platforms fold similar sharing layers into their own agent products.
The strategic pressure is clear. If Claude, Codex, Cursor, or another agent becomes the place where work is generated, each platform has an incentive to make its own artifact-sharing system sticky. Display.dev is taking the other side of that bet. Its homepage says teams can switch agent vendors while keeping URLs, versions, and comments in place. Tech.eu also reports that Display.dev is agent-agnostic and designed to support multiple assistants and models.
Ilves told Tech.eu that Display.dev started as "an elegant solution to one small problem" before the team saw a larger need for agent knowledge-work infrastructure. His sharper line was about adoption inside companies: "the productivity that agents produce today hasn't yet transferred into most companies." That is the commercial opening Display.dev is funded to test. The agents can make the work. Ilves and Rannaberg are betting companies will pay for the layer that lets people actually use it.