3.4×more qualified conversations than their old webinar Q&AShipiroo
Presentation maker AI assistants can drive — ask in chat, present live
DeckSlide is a presentation maker AI assistants can operate directly. Connect Claude, ChatGPT or any other MCP client to your account once, and the deck gets built where you are already typing — in the assistant’s own chat window, into your DeckSlide workspace. You open the studio afterwards to a deck that already exists, animate it properly, and present it live to a room that joins with a code.
- No card required
- Attendees install nothing
- Runs in any browser
Built for rooms you cannot afford to lose
- Hosted in the EU
- Imports your .pptx
- Audience joins with a code
Where does your team lose the most time?
Private threads from this room
3 openDoes this work for a 40-person team? I would like to see it with our own deck.
Waiting on youCan you go back to slide 4? I want to send that number to my manager.
You repliedSharing this with our design lead. What is the fastest way to try it?
Details shared+22%of attendees still in the room at the final slide
One address, one approval, no API key
You copy one address from Settings → Connections and paste it into your assistant as a custom connector. There is nothing to install and no API key to create — DeckSlide registers your client itself the first time it connects, so any field asking for a client secret is left empty. What follows is the part worth reading rather than clicking through: an approval screen naming the account the connection will act as, precisely what it will be able to do, and the host the access is going to. If that host is not the assistant you started from, press Cancel. Afterwards every connected client sits in Settings with its last-used time and a Disconnect button that bites immediately.
Real keyframes, not a tasteful fade
A lot of tools sell “AI animated presentations” and deliver a crossfade. Ask this one for motion and it writes actual keyframes: position, size and rotation over time, along a path you can curve, with the layer turning to face its direction of travel if that is what the shot needs. It will morph one shape into another as it moves, fade a layer through a gradient mask, and use the letters of a headline as the window a picture shows through — the poster trick people usually buy a motion designer for. Titles can take extruded depth, an outline or a block shadow in the same breath. And none of it is a black box: it is the studio’s own timeline, so you open the deck and drag the path, retime a key or change the easing yourself. The assistant cannot bring in an SVG file of your own as a layer; import that and it will animate it from there.
What it costs, precisely
Building, reading and editing decks is free, and so is everything to do with media you already own: listing your library, placing a picture or a track, importing one from a link, uploading a file from your machine. The single thing that costs is generated media — an image, a video clip or spoken narration made by AI — which spends your workspace’s token balance at the same rates as generating from the studio. The assistant can read that balance first and tell you what a request will cost, and “use my own images, generate nothing” is a constraint it will hold to. On the other side of the connection, your Claude or ChatGPT plan is billed by them under whatever you already pay.
The deck cannot leave — it is delivered live
One thing to know before you build your quarterly review here: nothing comes out as a file. No PowerPoint, no PDF, no MP4, no download, and no public link to the deck. The way a second person sees it is by joining your live session — they type an eight-character code or scan a QR, pick a display name and they are in, with no account and nothing installed. What they see is frozen at go-live, so a nervous edit mid-talk cannot change the slide under them. For an investor pitch or an all-hands that is a better shape than a file. For a leave-behind document it is the wrong tool, and we would rather you knew that on this page than after the work.
Questions, answered honestly
Which AI assistants can drive it?
Any client that supports remote MCP servers. Claude and ChatGPT both do — custom connectors are a paid-plan feature on their side, and on business workspaces an administrator may need to enable them — as do Cursor, Zed and a growing list of others. The steps are identical everywhere: paste the address, sign in, approve.
Is this really a free AI presentation maker with animation?
Building and editing with the assistant is free, the studio’s animation tools are free, and the assistant genuinely animates — real keyframes, curved motion paths, shape morphing and reveal masks, all editable afterwards. Only AI-generated media (images, video, narration) spends your workspace tokens.
Can I share the finished deck with someone as a link?
Not as a deck. A public deck URL and an embed both exist once you publish. An export never does. The shareable thing is a live session: you send a join code or QR and people watch you present it. That is the honest answer, and it is the whole design rather than a missing feature.
Can the assistant see my existing decks?
Yes — read access is part of what the approval screen lists, including speaker notes, across every workspace your account can reach. That is exactly why the screen names the account and the host, and why Settings keeps a Disconnect button next to every client.
Build the deck, then present it live.
Free to start. Your audience joins with a code — no account, no install.
Connect your AI assistant in Settings and ask it to build the first deck