Ott Ilves (@ottilves) header
Ott Ilves (@ottilves) profile picture

Building https://teapaw.org/t.co/Pt1mmKYfxI – the docs collaboration workspace for humans and agents. Father, sailor, runner and cyclist.

Joined November 2025
Photo by Ott Ilves (@ottilves)Photo by Ott Ilves (@ottilves)Photo by Ott Ilves (@ottilves)Photo by Ott Ilves (@ottilves)Photo by Ott Ilves (@ottilves)Photo by Ott Ilves (@ottilves)Photo by Ott Ilves (@ottilves)Photo by Ott Ilves (@ottilves)Photo by Ott Ilves (@ottilves)Photo by Ott Ilves (@ottilves)
Ott Ilves retweeted
Three new things in Display: 1. Annotate – pin a comment to the exact spot on a doc 2. Templates – start your next doc from one your team already uses 3. Header visibility – choose if docs open with the header open or collapsed
1
3
1
4
150
Ott Ilves retweeted
Companies are laying off thousands due to AI productivity gains. But why haven’t their products got any cheaper? Paying every license per seat for the whole toolstack teams use is not the elegant solution. To symbolise this, we staged a large chair in central London.
2
1
67
This is the biggest launch in the history of Display. As part of building this, we rewrote the whole underlying engine to support the much richer experience where you, your teammates and agents can work together.
Just 6 months ago, Display didn't yet exist. It was just an idea in our heads. So we built the small product and it's now used by more than 4000 people around the world. That's the @slashdisplay you know. Today, we're launching Display 1.0, the next phase and it's live on Product Hunt. We're introducing so many new things today, it's basically a whole new product. So our pitch is this: create docs and apps with any agent, in one workspace. Think of this as the GitHub for all your company's knowledge work, or an agent-native replacement for Notion. What's new in Display 1.0 > Spaces and Folders – give each team its own Space, and each client its own shared folder and decide who can see what. Organise work into Folders inside Spaces. >Storage – store and access any file type in your workspace. HTML, Markdown, PDF, images, videos – whatever you and your agents create and need storage for. > Storage – store and access any file type in your workspace. HTML, Markdown, PDF, images, videos – whatever you and your agents create and need storage for. > New Access Controls – introducing a more familiar Google Workspace-styled permissions matrix with editor, contributor, commenter and viewer roles. > User Groups – put people into groups and share any Space, folder or doc with the whole group at once. > External contributors – updated access controls also let you give commenting and editing access for docs to external guests. Combined with Spaces and Folders, you can now create dedicated client spaces where you work together on any agent or human created doc with your partners. > Internal Knowledge Base – the new document storage, spaces, and updated search let you create an internal knowledge base of all your company's know-how, providing context for everything your agents do next. We have internal apps and data connectors working in private beta, available shortly to everyone. So my ask is this - please head over to Product Hunt and give us an upvote!
2
1
6
116
Ott Ilves retweeted
One cool use case for visualize is direction brainstorming. Visualize is your companion for building single-purpose landing pages, prototypes and mockups. Full tutorial here: youtube.com/watch?v=BXLOGFNU…
1
3
3
96
Ott Ilves retweeted
Proud to announce usevisualize.ai It's a design skill we've been using a lot internally and decided not to gatekeep anymore. Find setup instructions and examples on the link. Hope you'll have as much fun using it as I do!
Our internal visualization companion goes public! How to access it? 1. Go to usevisualize.ai 2. Copy the setup prompt 3. Let your agent do the rest Now start designing beautiful on-brand HTML docs without the AI-default slop.
2
3
113
Ott Ilves retweeted
Replying to @slashdisplay
@slashdisplay visualize skill helped me become a product designer. Just kidding but I'm proud of the progress indeed. I also showed the next steps, so how to actually go from prototyping to getting it live. Cheers to my partners in crime Karola and @claudeai
1
3
8
168
Ott Ilves retweeted
This is the exact way builder and designers use display.dev to take feature ideas to prototypes to shipped features. Start building and prototyping today.
We solved cross-agent collaboration in prototyping. Designers and builder prototyping is one of the most popular use cases for display.dev. Exploring and collaborating different design options for features -- with different agents -- is also how @carlrannaberg and I work together every single day. So I created a run-through of how I went from an idea to a prototype to a shipped feature, using my agent while iterating with comments. The set-up is simple: > My agent has access to my repository, so it knows my product context > I use the prototype skill - a set of instructions that add various testing options to the prototypes the agents create - to explore different designs for the feature. > We iterate on the prototype version I liked with comments, updating the design until we were ready > I ask the agent to create a final spec and develop the feature, using our building and deployment workflow The agents do the development, display.dev is the place where review and collaboration happens. How do you run design exploration for your product?
1
2
79
We built a prototype skill to explore new UI directions using display.dev It has sped up UI work quite a bit. Below Ott is demoing one real use case end-to-end.
We solved cross-agent collaboration in prototyping. Designers and builder prototyping is one of the most popular use cases for display.dev. Exploring and collaborating different design options for features -- with different agents -- is also how @carlrannaberg and I work together every single day. So I created a run-through of how I went from an idea to a prototype to a shipped feature, using my agent while iterating with comments. The set-up is simple: > My agent has access to my repository, so it knows my product context > I use the prototype skill - a set of instructions that add various testing options to the prototypes the agents create - to explore different designs for the feature. > We iterate on the prototype version I liked with comments, updating the design until we were ready > I ask the agent to create a final spec and develop the feature, using our building and deployment workflow The agents do the development, display.dev is the place where review and collaboration happens. How do you run design exploration for your product?
1
2
78
This is how work looks like now. For better or worse. For context, I asked the agent to create an synthesis of multiple research docs. It uses our internal writing skill that reduces the slop, Minto Pyramid skill to shape how the information is laid out, visualize skill to visualize it as HTML document and finally display.dev CLI to publish to our company workspace and give back a shareable URL. #ai #agents #llm #codex
1
42
Ott Ilves retweeted
Just shipped: 1. Artifacts now show who else has them open (anonymous viewers never appear) 2. Cmd + k search for every dashboard page and artifact, and 3. Add your profile photo, visible across the product See more -> display.dev
1
1
48
Ott Ilves retweeted
This is how you do client reporting in one minute instead of one hour. Your agent pulls the numbers – ad accounts, analytics, email, Slack – writes the update and publishes with display.dev. Share it with your client, who sign in with a one-time password. Done.
2
6
51
Ott Ilves retweeted
New in display.dev: verified email domains. Add every domain your company has – the acquired brand, the subsidiary, the legacy alias – and prove each one with a DNS record. Everyone on a verified domain can join your org and open company-wide artifacts, no invites.
1
2
1
2
120
Ott Ilves retweeted
Display.dev raises €470K to power document collaboration for AI agents tech.eu/2026/07/07/displayde…
3
1
2
311
Ott Ilves retweeted
One feature, two agents, one URL. Carl published v1 from Codex. Claude Code iterated to the version that shipped – every hand-off attributed in the audit log. A 3-minute walkthrough of a real display.dev workflow:
1
2
1
2
162
Ott Ilves retweeted
14 versions, 2 agents, 1 URL – how to use display.dev in practice > Carl's Codex published v1 and v2 of our signup modal > Ott made v3 and v4 with Claude Code > Carl took it from v5 to production with Claude Code Audit logs show which agent published each version
2
3
113
Ott Ilves retweeted
Just a reminder. Stay agent-agnostic. The best models and agents change, your important work should always be accessible.
1
2
30
Ott Ilves retweeted
Introducing Visualize A free skill for on-brand HTML with your coding agent. Agent pages all look the same. This fixes it: → Templates + design systems → Iteration guidelines ("make it bolder") → Works with any agent GitHub: github.com/display-dev/visua…
1
1
55
Ott Ilves retweeted
New customer story: Indigo Engineering display.dev/customers/indigo…
1
1
54
Ott Ilves retweeted
Agent identity is here. When an agent edits your doc, replies to a comment, or resolves a thread, you should know exactly who did it and on whose behalf. Full loop in the walkthrough.
1
1
36