An automatic slideshow maker turns data you already keep, such as a listing feed, a product CSV or a website’s sitemap, into one finished video per row or page without hand editing. It should choose the photos and their order, write the script, record the voice, add captions and export each format, then leave a person to approve the result. Videngine does this for businesses, rendering narrated, captioned 9:16 and 16:9 slideshows in unattended batches.
Transcript
An automatic slideshow maker should take everything an editor decides and pull it straight from your data. That means choosing which photos to use, setting their order and timing, writing a script from the facts in each record, recording a voice, adding captions, and exporting every format you need. What it should not decide is whether a video goes live.
Every finished video waits for a human to review it first. The source for each video is one record: a title, a set of facts, photo URLs in order, a status, and a link. Before anything renders, quality gates check that photos are sharp enough, that key facts are present, and that the item is still available.
Videos can be queued in batches from a CSV or a sitemap, running unattended while a person reviews each result. When a price or status changes, the same record re-renders as a new version.
What an automatic slideshow maker should decide for you
Automatic should mean that every per-video choice an editor would make comes from your data, leaving approval as the only job per video.
For a listing or a product, that is six decisions:
- Photos: from the page or the row, with soft ones flagged, not quietly used.
- Order and timing: a scene plan setting which photo appears when, and for how long.
- Script: written from the record’s facts, within your pronunciation, banned-phrase and length rules.
- Voice: a house voice, another synthetic voice, or audio you upload.
- Captions: transcribed from the finished audio, so screen and speech match.
- Formats: 9:16 and 16:9 cuts, plus 1:1 and 4:5 when needed, up to 1080p.
What it should not decide is whether a video goes out: in Videngine every video is reviewed in the studio first, where you can accept the script, edit a line or supply your own. When comparing any AI slideshow maker or auto video maker, ask which of these it takes from your data. The slideshows hub covers the format itself.
| Question | Made by hand | Made automatically |
|---|---|---|
| Time per video | Every step, every video: choose, order, write, record, caption, export | Review only; a one-minute 1080p video typically renders in a few minutes |
| Consistency | Varies with the editor and the day | Same template, brand tokens and voice rules every time |
| What needs a person | Everything | Template set-up once, flagged photos and missing facts, approval |
| A price change | Reopen the project and re-edit | Update the record and re-render |
| Two hundred listings | Two hundred edits | One unattended batch, then two hundred reviews |
The data each video needs
Each video needs one record holding a title, a few facts, its photos in order, a price or status and a link; nothing the narration says should come from anywhere else.
Any slideshow generator that works from data needs roughly these fields, and a missing one shows in the video.
| Field | Example | Drives | If missing |
|---|---|---|---|
| ID | PX-1042 | File names and versions | Re-renders cannot be matched to listings |
| Title | Three-bedroom terrace, Hove | Opening card and first line | Hold the row |
| Facts | 3 bedrooms; 2 bathrooms; garden | The body of the script | Nothing specific to say |
| Photo URLs, in order | photo_1 to photo_8, lead first | Photo scenes and their sequence | The row fails the photo check |
| Price or status | 450000 GBP; Under offer | Closing line; when to re-render | Omit the price on purpose, not by accident |
| Link | The listing or product page | Where the viewer goes next | No destination |
A CSV header for a property feed might look like this. The names are illustrative; what matters is one row per video, with photos in display order.
id,title,price,currency,status,facts,url,photo_1,photo_2,photo_3,photo_4,photo_5,photo_6,photo_7,photo_8Quote any value containing a comma, such as "Three-bedroom terrace, Hove", as RFC 4180 describes for CSV files. If the facts already sit on a public page, skip the spreadsheet: Videngine reads the page in a headless browser and keeps its text, images and structured data. The guide to bulk video creation from a CSV covers row validation and job tracking.
Quality gates before anything renders
Check photos and facts before a batch is queued, because a render is only as good as the row it came from.
| Gate | Rule | If a row fails |
|---|---|---|
| Photo count | Four to five seconds a photo: six to eight for a 30-second cut | Hold it, or accept a shorter video |
| Photo grade | By the original’s long side: sharp at 1,600 px or more, ok from 1,000 px, low below that | Replace low photos, or keep them out of the lead |
| Missing facts | Title, key facts and price or status all present | Fix the source; the script should not work round a gap |
| Status | Only items still for sale or to let | Skip sold and withdrawn items |
Videngine handles part of this. It asks a site for original photos before resized copies. It enlarges photos under 1,920 px on the long side with a Lanczos filter and a light sharpen, never more than 3×, so a 600 px thumbnail cannot fill a 1080p frame sharply. It turns photos the right way up from their orientation data and flags soft ones in the studio.
A photo that does not fit the frame sits over a blurred fill of itself, so landscape photos work in 9:16 without losing the subject; the vertical slideshow guide shows how.
From a sitemap or CSV to approved videos
Set the brand once, choose and check sources, prove one video, then batch the rest and review each result.
- Set up the workspace. Pick the business type so the studio’s wording fits, then set typefaces, colours, logo, safe areas and motion once, plus voice rules: a pronunciation lexicon, banned phrases and a target length.
- Choose the sources. Upload a CSV or JSON with one record per video, or have Videngine read your sitemap (an XML file with an entry per URL, under the sitemaps.org protocol) and pick the pages you want.
- Run the gates. Photo counts and grades, missing facts, status.
- Prove one video. Render one representative record, check the script and scene plan, and approve the template there.
- Queue the batch. One row or page becomes one video, usually with a 9:16 and a 16:9 cut. Batches run unattended.
- Review every video. Approve it in the studio, or fix the record and rebuild it as a new version.
- Deliver. Download MP4s, or send them to a Google Drive folder you own, a connected YouTube channel, or the source page on a GitHub-connected static site.
The steps are the same for an agent’s forty listings and a marketplace’s full feed; see real estate listing video, video for marketplaces and, for catalogues, product video automation.
Keeping videos current when prices change
Re-render a video whenever a fact it states changes, and keep the earlier version rather than overwriting it.
Typical triggers are a new price, a status change such as under offer or sold, and new photos. Compare each video’s ID, stated price, status and version with your feed before each batch. An accurate sitemap lastmod, which the protocol defines as the date of last modification of the page, makes a cheap first filter.
In Videngine, rendering the same video again in the same month is free, and the video ledger saves a rebuild as a new version, such as _v2, beside the original. On a website page published from Videngine, republishing replaces the video in place and Remove takes it out in one commit. On YouTube, decide what happens to the earlier upload.
If prices move daily, leave the price out and let the page carry it. A video that states no price cannot state the wrong one.
Cost per approved video
On Videngine’s published launch prices, a plan used in full costs between $1.58 and about $0.80 per video, and rendering the same video again in the same month costs nothing extra.
| Plan | Per month | Videos included | Per video, plan used in full | Extra videos |
|---|---|---|---|---|
| Starter | $79 | 50 | $1.58 | $1.50 each |
| Growth | $249 | 250 | About $1.00 | $1.00 each |
| Scale | $599 | 750 | About $0.80 | $0.80 each |
A video is one finished render up to 2 minutes; a longer one counts as two. Custom plans start from 1,500 videos a month; seats and brands are included. Prices are in US dollars, as on the pricing page. Videngine is coming soon: choosing a plan joins the launch list, with no charge today.
Batch runs from a CSV or JSON file start on the Growth plan; Starter makes one video at a time from a link, text or your own script. Example: 180 listings a month, one video each, run as a batch: Growth covers all 180 for $249, about $1.38 a video. Divide by approved videos, not attempts, then add the review minutes per video, which automation does not remove.
When Videngine is the wrong tool
Videngine is built for many narrated videos from business data; for personal slideshows, music-only edits or animated stills, another kind of tool fits better.
- One or two personal slideshows. For a birthday or a wedding, a free consumer tool fits: your phone’s photo app, Google Photos, iMovie or Canva.
- Music-only edits. Slideshows set to a song, beat-synced cuts, collages and grids need a consumer editor, a music-led slideshow tool or an AI montage maker. Videngine adds only an optional quiet music bed under narration; see slideshows with music.
- Animating stills. Turning a photo into new motion is AI image-to-video. Videngine uses a slow zoom and a short fade and never invents movement; image-to-video AI vs a slideshow compares the two.
- A one-off film or another language. A crafted brand film needs a timeline editor. Videngine narrates in English only.
For a single video made by hand, our guide on how to make a slideshow video covers photo choice, order and timing.
Questions about automatic slideshow makers
What does an automatic slideshow maker do?
It turns photos and facts you already keep into finished slideshow videos, many at a time, without hand editing. It takes photos, order, script, voice, captions and formats from your data, and leaves a person to approve each video.
Can a video slideshow generator make hundreds of videos from a spreadsheet?
Yes, if each row holds everything one video needs: a title, facts, photo URLs in order, a price or status and a link. Videngine turns each CSV row into one video, runs the batch unattended, and holds every video for review in the studio.
How many photos does an automatic slideshow need?
Enough to fill the length at four to five seconds a photo: six to eight for a 30-second video. Size matters as much as count. Videngine grades each photo by its original long side, sharp from 1,600 px and ok from 1,000 px, and flags soft ones.
Is an AI slideshow maker the same as an AI video generator?
No. Videngine’s AI slideshow maker uses your real photos; AI writes the script and the voice, not the pictures. An AI video generator invents footage or animates stills into new motion, a problem when the video must show the actual property or product.
Can I add music to an automatic slideshow?
In Videngine, an optional quiet music bed can sit under the narration. It does not make music-only slideshows set to a song or beat-synced edits. For those, a consumer editor or a music-led slideshow tool is the better fit.
Choose your next step
Start with one representative record, then decide how far to automate.
Build your video workflow.
Bring one representative listing or product record and the formats you publish. Prove the slideshow on that record, then run the rest as a batch.
Join the waitlist ↗Published by Videngine, a video rendering product built and run by Wall & Fifth. External links refer to the original documentation or product descriptions; examples are illustrative unless stated otherwise.