How do I add images to slides in Rideful?

In the editor, paste an image with Cmd/Ctrl+V, drag files onto the slide, or click Insert Image in the toolbar to search stock photos or upload from your device. In the deck's chat, paste an image or give the agent a public HTTPS link and ask it to place it, or ask for a picture you do not have and the agent finds or generates one. From an MCP client, use element_create_image with any public HTTPS URL. Rideful stores the image permanently either way. Supports PNG, JPEG, GIF, and WebP up to 10MB.

Can the Rideful agent find a real photo for my slide?

Yes. Ask for a picture of something real and the agent searches Pixabay for generic scenes and Wikimedia Commons for named people, places, buildings, artworks and products, picks one from the thumbnails and places it. A named person or place is never generated. The credit goes in the picture's alt text, which the PowerPoint export keeps, and in a small caption when the licence requires attribution. Over MCP, image_search returns the same candidates.

Can I search stock photos in the Rideful editor?

Yes. Insert Image in the toolbar opens a search of Pixabay's free stock photos, with Any, Landscape and Portrait filters and More for the next page. One click places the photo on the slide or layout, centered and fitted inside 80% of it. Its credit, "Image by Jane Doe from Pixabay", goes in the alt text, which survives duplicating and undo and is written to the PowerPoint export.

How does image sizing work in Rideful?

Set width only to auto-calculate height (preserving aspect ratio), height only to auto-calculate width, both to fit within a bounding box, or neither to auto-fit at up to 80% of the slide canvas. Images never upscale beyond their original dimensions.

Can I crop images in Rideful?

Yes. Double-click an image to enter crop mode, drag a handle to move an edge or drag the picture to choose what shows, and press Enter to keep it. The image toolbar's Crop menu also cuts selected images to 1:1, 4:3, 3:4, 16:9 or 9:16, and Reset crop shows the whole picture again. The agent crops on request, and over MCP the crop property takes percentage values (0-95) trimmed from each edge (left, top, right, bottom). Cropping is non-destructive — the original image is preserved and you can adjust or remove the crop at any time.

What image formats does Rideful support?

Rideful supports PNG, JPEG, GIF, and WebP images up to 10MB, in the editor and over MCP alike. SVG is not supported. Images added from a URL must additionally use HTTPS (HTTP is rejected for security) and the URL must be publicly accessible.

How Do I Add Images to Slides?

Paste, drag, or upload an image in the editor, or search stock photos from its toolbar; paste one into the deck's chat and ask the agent to place it; give an agent a public HTTPS URL; or ask the agent in the chat for a picture: it finds a real photo, credited, or generates a new one. Either way Rideful stores it permanently, sizes it to the slide, and exports it faithfully.

In the editor you paste, drag, or pick a file, or search Pixabay's stock photos from the toolbar. In the deck's chat panel you paste, drag, or pick an image into the composer, or give the agent a public HTTPS link, and ask it to put the image on a slide; or ask for a picture you don't have, and it finds or generates one. Over MCP an assistant finds photos with image_search and adds one from a URL with element_create_image. Once an image is on the slide, the agent will move it, resize it, layer it, crop it, mask it to a shape, give it a shadow, or set its alt text.

In the Editor

Three ways in, all equivalent once the image lands:

Paste — copy an image anywhere and press Cmd/Ctrl+V with the slide focused. The first image on the clipboard is added, centered.
Drag and drop — drop image files onto the slide. They land where you drop them; several at once are staggered slightly so they don't hide each other.
Insert Image — the image button in the toolbar opens a stock photo search (below), with Upload from device at its top for your own files. Upload accepts several files at once, and they are centered.

The image appears immediately at its final size and position while the upload runs in the background, so you can keep working. Once it finishes it is an ordinary element: select, move, resize, and undo it like anything else. If the upload fails the preview stays put with a retry.

From the Toolbar

Insert Image opens on a search of Pixabay, a free stock library. Type what the photo should show, such as “team whiteboard”, and the grid fills once you stop typing, or press Enter. Any, Landscape and Portrait narrow it by shape, and More under the grid adds the next page. Hover a photo to see who took it: the name opens the photo's page on Pixabay. The Images from Pixabay link stays at the bottom of the panel.

One click places the photo on the slide or layout you are editing, centered and fitted inside 80% of it, never blown up. A preview appears at once while Rideful stores its own copy; then it is an ordinary image, selected, that you move, crop, mask and undo like an upload. If storing it fails, the preview stays with a retry.

Its credit travels in its alt text: a description of the photo from its Pixabay tags, then “Image by Jane Doe from Pixabay”. Duplicating it, copying and pasting it, or undoing its deletion keeps it, and the PowerPoint export writes it as the picture's alt text. Pixabay photos need no visible caption.

Pixabay's licence bars misleading use, especially of photos with recognisable people: don't set a stock face beside a product claim as if that person were a customer.

In the Chat

Paste an image into the chat panel's composer, drag it there, or tap the image button beside Send to pick one (on a phone, from your photos or the camera). It uploads straight away. A message carries up to four images; a large photo is scaled down to fit the 5 MB limit. The agent sees them, so you can ask about one (“does this chart read well?”), ask it to put one on the slide, or ask for a slide that looks like it (“match this screenshot”): the designer building that slide then sees the image as a reference. It places the image on the slide you name, or the one you are on, sized to fit. A link works the same way: give the agent a public HTTPS URL to an image and ask it to add it.

An image pasted into the chat is kept for 30 days. Once the agent places it on a slide, the slide has its own permanent copy.

Generated by the Agent

Ask the agent in the deck's chat for a picture that does not exist yet — “a photo of a quiet harbour at dawn for the title slide”, “an abstract background in our blues behind the stats” — and it generates one and places it filling the space it chose on the slide. Paste a photo and ask for it restyled, and the generated picture follows it.

No text in the picture. Titles, labels and numbers stay ordinary text on top, so you can still edit, restyle and translate them.
Matches the deck. Pictures follow the deck's palette and one imagery style, so the slides look like a set.
Only when asked. The agent adds a picture when you ask for one or approve a plan that includes one, not to fill space. Icons, charts and diagrams stay native elements.
Takes up to a minute per picture, and counts toward your daily assistant usage.

Once placed, a generated picture is an ordinary image: move, resize, crop, delete or undo it like any other. Generation is in the deck's chat only; over MCP, find a photo with image_search or add one from a URL with element_create_image.

Found by the Agent

Ask the agent for a picture of something real — “a portrait of Max Wertheimer on slide 3”, “a photo of a team at a whiteboard filling the left half” — and it searches for an existing photo, picks one by eye from the thumbnails, and places it.

It searches two sources, both of which state a licence on every file. It never searches the open web, where a picture rarely says whether you may use it.

Pixabay — a free stock library, for generic subjects: people at work, places, objects, nature.
Wikimedia Commons — for a real, named person, place, building, artwork, event or product. A name such as “Eiffel Tower” finds the picture that leads its Wikipedia article.

The subject decides between finding and generating:

A named person, place or thing is always found, never generated. A generated “Max Wertheimer” would be a plausible stranger under a real name. When Commons has nothing that fits, the agent says so: paste or link a photo you have and it places that.
A generic scene is searched first, and generated only when nothing found fits the slide's colours or composition. A deck whose pictures are illustrations generates its scenes.
An abstract idea, a texture or a background in the deck's palette is generated.

A found photo fills the space the agent chose, cropped about its centre, and Reset crop in the Crop menu shows the whole photo. An old portrait on Commons can be small, so the agent places it at its own size rather than stretching it soft.

Credits travel with the picture:

Alt text describes the picture and ends with its credit: “Image by Jane Doe from Pixabay”, or “John Doe / CC BY-SA 4.0” from Commons. The PowerPoint export keeps it as the picture's alt text.
A caption beside the picture, in small body text, when the licence requires attribution (CC BY and CC BY-SA on Commons). Keep it if you move the picture. Pixabay photos and public-domain files need none.
In the chat, the search shows its candidates as thumbnails, each linked to its page, and a stock search carries an Images from Pixabay link. The step that placed the photo shows its credit, linked to the source, and stays in view after the turn's other steps fold away.

A found face or logo is never set beside a product claim as if that person or brand endorsed it. When a file carries a restriction, such as personality rights or a trademark, the agent tells you.

A search costs nothing and takes about a second, where a generated picture takes up to a minute and counts toward your daily usage. The agent makes at most five searches per turn. Once placed, a found photo is an ordinary image, stored in Rideful like any other.

Over MCP, image_search returns the same candidates, each with its thumbnail, licence and credit line. Place one with element_create_image, passing its url as sourceUrl and fit: "cover" with the box's width and height; the credit goes in altText, and in a caption when attributionRequired is true.

Supported Formats

PNG, JPEG, GIF, WebP
Maximum file size: 10MB
No SVG — SVG images are not supported

Both surfaces enforce the same formats and the same 10MB cap. The URL path adds one rule of its own: HTTPS only — HTTP is rejected for security.

Sizing and Positioning

You have full control over how images are sized and placed on the slide:

Set width only — height is calculated automatically to preserve the aspect ratio
Set height only — width is calculated automatically to preserve the aspect ratio
Set both — the image fits within the bounding box while preserving its aspect ratio
Set both with fit: "cover" — the image fills the box exactly, cropped about its centre (not with crop)
Set neither (auto-fit) — the image scales to fill up to 80% of the slide canvas without upscaling, and centers on the slide

Position with x and y in pixels from the top-left corner. If omitted, the image centers on the slide.

The editor applies the same auto-fit: an image larger than 80% of the canvas is scaled down to fit, a smaller one is never blown up, and it centers unless you dropped it somewhere specific.

Cropping

Double-click an image to crop it, or pick Crop from the Crop menu in the image toolbar or from the image's right-click menu. Crop mode shows the whole picture faded behind the part that stays:

Drag a handle to move that edge. The picture stays where it is; only the part that shows grows or shrinks, and an edge stops at the picture's edge. Hold Shift on a corner to keep the shape.
Drag the picture to slide it under the frame and choose what shows.
Press Enter, pick Done in the Crop menu, or click anywhere else to keep the crop. Escape or Cmd/Ctrl+Z cancels it.

The kept crop is one undo step. A rotated image crops along its own edges, and an image on a layout crops the same way in the layout editor.

The rest of the Crop menu applies to every selected image at once, in one undo step:

1:1, 4:3, 3:4, 16:9, 9:16 — cuts each image to that shape about the middle of what shows.
Reset crop — removes the crop: the image grows back to the whole picture, unstretched, where it sits. A picture the agent generated or found is cropped to fill the space it was given, so this shows all of it.

In crop mode they change the frame instead, and keeping the crop saves them; after a ratio, drag the picture into place.

The agent crops too: ask in the chat (“crop 20% off the left of that photo”), or set crop over MCP, trimming from each edge in percent (0–95):

crop.left — percentage to hide from the left edge
crop.top — percentage to hide from the top edge
crop.right — percentage to hide from the right edge
crop.bottom — percentage to hide from the bottom edge

Cropping is non-destructive — the original image is preserved and you can adjust or remove the crop at any time. The element's width and height represent the visible area after cropping.

Crop constraints

The total crop on each axis must leave at least 5% visible: left + right ≤ 95 and top + bottom ≤ 95.

Masking

A mask clips the picture to a shape: a round headshot, a photo with rounded corners. Select one or more images and pick a shape from Crop to Shape in the image toolbar, PowerPoint's name for it: Rounded, Ellipse, Diamond, Triangle, Left or Right arrow, Hexagon or Star, with Rectangle removing the mask. The Corner Radius slider beside it rounds the corners, and hides once another shape is picked. Or select one image and pick a Mask in the properties panel:

None — the plain rectangle every image starts as
Ellipse — an oval that fills the image's box
Rounded rectangle — with a Corner radius in pixels, starting at a sixth of the shorter side; a radius of 0 removes the mask

The mask fills the image's box, so an ellipse on a 16:9 photo is an oval. For a round headshot, make the box square first: pick 1:1 in the Crop menu, then Ellipse in Crop to Shape, then double-click the image and drag the face into the circle. Or ask the agent for a round headshot and it does the same. Crop and mask combine as in PowerPoint: the crop picks which part of the picture shows, and the mask clips the box. The image's fill and outline follow the mask. Every change is one undo step.

Over MCP, set mask on element_create_image or element_bulk_update: { shapeType: "ellipse" }, or { shapeType: "rectangle", cornerRadius: 24 }. null, or a rectangle without a radius, removes it.

A picture cropped to a shape in an imported .pptx stays an image with its mask, whether an ellipse, a rounded rectangle, a polygon or a custom cut-out, and exports with the same outline. The panel shows a mask it does not offer by name, and None removes it.

Shadows

A shadow lifts a picture off the slide. Select one or more images and pick Shadow in the image toolbar, or select one and pick it in the properties panel:

None — no shadow
Soft — a light shadow just below the picture
Lifted — a wider, deeper one that floats it

The shadow follows the picture's visible pixels, as in PowerPoint: a cut-out on a transparent background, such as a generated object, casts its own outline, and a photo casts its rectangle. A mask, a crop, a fill and an outline shape the shadow too. A full-bleed photo needs none.

Over MCP, set shadow on element_bulk_update: 'soft', 'lifted', or 'none' to remove it. element_create_image takes no shadow, so add the image first. In the chat, ask for one (“give the product shot a soft shadow”).

A picture's shadow in an imported .pptx keeps its own values, and the menu shows it as Custom until a preset replaces it.

Code Example

// Add an image with automatic sizing and centering
element_create_image({
  deckId: "deck_abc",
  slideId: "slide_1",
  sourceUrl: "https://example.com/photo.jpg"
})

// Add an image with specific width (height auto-calculated)
element_create_image({
  deckId: "deck_abc",
  slideId: "slide_1",
  sourceUrl: "https://example.com/photo.jpg",
  x: 60,
  y: 120,
  width: 500
})

// Fill a 640×720 box exactly, cropped about the centre
element_create_image({
  deckId: "deck_abc",
  slideId: "slide_1",
  sourceUrl: "https://example.com/wide-photo.jpg",
  x: 0,
  y: 0,
  width: 640,
  height: 720,
  fit: "cover"
})

// Add an image with cropping (trim 20% from the left)
element_create_image({
  deckId: "deck_abc",
  slideId: "slide_1",
  sourceUrl: "https://example.com/wide-photo.jpg",
  x: 0,
  y: 0,
  width: 640,
  height: 720,
  crop: { left: 20, top: 0, right: 0, bottom: 0 }
})

// A round headshot from a 16:9 photo: crop the sides to a square, then mask
element_create_image({
  deckId: "deck_abc",
  slideId: "slide_1",
  sourceUrl: "https://example.com/portrait-wide.jpg",
  x: 80,
  y: 160,
  width: 240,
  crop: { left: 21.875, top: 0, right: 21.875, bottom: 0 },
  mask: { shapeType: "ellipse" }
})

How It Works

From the editor: the file is checked for format and size, its dimensions are read to compute the display size, a preview goes on the slide straight away, and the upload to cloud storage runs behind it. The finished element goes onto the undo stack like any other edit. A stock photo is fetched from Pixabay by Rideful's server with the same format and size checks as a URL, then placed the same way.

From a URL:

URL validation — Rideful checks that the URL is HTTPS and points to a supported image format
Server-side fetch — the image is downloaded from the source URL (10-second timeout, up to 3 redirects followed)
Size check — the file must be under 10MB
Permanent storage — the image is stored in Rideful's cloud storage, so it remains available even if the original URL goes down
Dimension detection — original width and height are read for aspect ratio calculations
Element creation — an image element is added to the slide with the calculated size and position

Common Issues

In the editor:

“Unsupported image format”: the file is not PNG, JPEG, GIF or WebP. SVG and PDF are rejected here too.
“Image exceeds 10MB”: compress it or export at a smaller resolution.
“Upload limit reached”: more than 50 uploads in a minute, stock photos included. Wait a moment and retry.
An upload that failed: the preview stays on the slide with a retry — it is not lost, and nothing was written until it succeeds.
“Stock search is busy”: Pixabay is limiting searches for the moment. Try again in a minute.
“Stock photos aren't available right now”: stock search is switched off. Upload from device still works.
“Couldn't add photo” after the panel sat open for a long time: Pixabay's links expire after about a day. Search again and click the photo from the fresh results.

From a URL:

“Source URL must use HTTPS”: Only HTTPS URLs are accepted. If your image is on HTTP, re-host it on HTTPS or use a CDN that supports it.
“Image exceeds 10MB limit”: Compress the image or use a smaller resolution before adding it.
“Request timed out”: The source server took too long to respond (limit is 10 seconds). Try a different URL or a faster host.
SVG images: SVG is not supported. Convert to PNG or JPEG before adding.
Image not showing in PowerPoint export: images are copied into Rideful's own storage however they were added, so they always export — even if the original file or URL is gone.

Styling — Colors, fonts, sizes, and shape formatting
Canvas Size — Slide dimensions and coordinate system for positioning
Create with AI — Create full presentations including images using natural language
Export to PowerPoint — Images are embedded directly in the .pptx file