Most teams never do this. They generate fresh each time, fight the same font mismatch, and ship something that looks like it came from three different companies.

Why start from a saved project instead of the generator?

The generator optimizes for getting video out fast, not for consistency. Its default caption style, transition timing and logo placement reset on every new project.

A saved base project carries your decisions forward. Duplicate it, paste new captions, swap footage — the typography and motion stay fixed. That is the whole trick, and it takes about 40 minutes to set up once.

How do you build the base project?

1. Create a project manually, not from a URL. Empty canvas, single scene, 16:9.
2. Set typography first. One heading font, one body font. Lumen5 exposes limited font controls — pick from what is available rather than fighting it.
3. Lock a two-color palette. Primary for captions, accent for CTA buttons. Keep contrast above 4.5:1; overlay text on busy B-roll fails otherwise.
4. Build a 2-second intro block. A static logo on a solid color renders faster and looks more deliberate than an animated stock intro.
5. Build a 4-second outro. One line of text, one CTA, your URL. This is the only place the logo appears.
6. Add a lower-third scene pattern. Duplicate it eight times as placeholders in the timeline.
7. Save and name it BLANK - 16:9 - v1. Duplicate that for every new video, never edit the original.

What should the template enforce, and what should it leave open?

Enforce anything viewers notice without thinking: font, color, logo position, caption alignment, safe margins.

Leave open anything the content dictates: scene count, clip choice, whether there is a talking-head segment.

A useful test: if a viewer could tell two videos came from the same company without seeing the logo, the template is doing its job.

Is the Professional tier required?

For a real library, yes. Starter at $19/month lacks the brand kit, which means your saved fonts can still be overridden by the generator’s defaults on some export paths.

Professional at $59/month adds custom font upload and a saved palette. On a team shipping 10 videos a month, that is roughly $4 per video above Starter. Cheap if it removes an hour of alignment work.

Where this breaks down

Templates calcify. Six months in, every video has the same three-transition rhythm and the channel starts feeling automated. The fix is quarterly review — retire one pattern, add one.

Also: aspect ratio is per project. You need a separate base for 9:16 and 1:1, which triples setup to roughly two hours. Worth it once, annoying to maintain.

Key Takeaways

• Duplicate a saved base project rather than generating from a URL each time.
• Professional tier at $59/month is needed for custom fonts and locked palettes.
• Setup takes about 40 minutes per aspect ratio; expect roughly two hours for three ratios.
• Enforce typography, color, logo position; leave scene count and footage flexible.
• Review and retire stale patterns quarterly or the output starts feeling automated.

The interesting part is not the automation. It is what you have to undo after the automation runs.

What Lumen5 actually does with your article

The engine pulls the article, splits it into sentences, scores them for standalone meaning, and keeps the top ones as scenes. Each scene gets a text overlay and a suggested stock clip matched to keywords in that sentence.

The scene count depends on length. A 1,200-word post typically yields 12-18 scenes at default density. You can raise or lower density with a slider, which changes the number of scenes but not the source selection quality.

How do you run the paste-a-link workflow?

1. Log in and hit “Create video”, then “Blog to Video”. Paste the full post URL, not a preview link — paywalled or JS-rendered pages return empty.
2. Set the ratio before generation. 16:9 for YouTube and LinkedIn, 1:1 for feed, 9:16 for Shorts. Changing later re-flows every caption.
3. Let it generate, then stop. Do not start editing scene one. Read the whole storyboard first.
4. Delete ruthlessly. Aim to cut the scene count by a third. Generated videos feel long because every sentence gets a scene.
5. Replace stock clips that say something wrong. This is the step people skip. Keyword matching produces literal, often absurd B-roll.
6. Add your logo to the outro only. A persistent watermark plus Lumen5’s brand mark looks amateur.
7. Export 1080p, watch muted, then watch with sound. Captions must carry the message alone.

Why is the auto-selected footage usually wrong?

Keyword matching has no idea what a sentence means. A sentence about “cutting your cloud bill” gets a shot of scissors on paper. A line about “churn” gets a literal churn.

The workaround is to treat generated clips as placeholders. Lumen5’s library is large — millions of clips — but relevance is a semantic problem, and this is a lexical matcher. Budget 10-15 minutes per video purely for clip replacement.

If you have your own footage, upload it first and the matcher will pull from your library before stock. That single step halves the cleanup.

What does Lumen5 cost in 2026?

Annual billing knocks about 20% off. The jump from Starter to Professional is mostly about voiceover — Starter gives you no text-to-speech minutes on some plans, so check the current tier sheet before assuming.

Who is this actually for?

Content marketers repurposing a blog archive. If you have 80 posts and no video team, batching 10 videos from old articles is a realistic afternoon.

Not for anyone selling a visual product. Stock footage cannot show your app’s interface, and Lumen5 has no screen recording.

Key Takeaways

• Lumen5 turns a URL into scenes automatically; a 1,200-word post gives 12-18 scenes.
• Free tier exports 720p with a watermark; Starter is $19/month for 1080p.
• Generated stock footage is a keyword match, not a semantic one — expect to replace most clips.
• Uploading your own media first makes the matcher prefer it over stock.
• One ratio per project; re-export for each platform.