A colorizer that behaves less like a button, more like a workflow

Palette.fm positions itself as a “new generation” AI colorization tool: quick enough for casual use, but structured for repeatable output.
The core flow is intentionally simple:

1. Upload your photo (including drag-and-drop for multiple images).
2. Pick a filter to set the overall look and mood.
3. Download or edit—either save the result or customize the colorization via keywords.

The practical difference is that Palette doesn’t stop at “here’s the colorized image.” It encourages iteration—try multiple looks, then tighten the result
with text-based guidance.

Why people pick Palette: realism, detail, and the ability to “nudge” the model

Palette highlights three recurring reasons users stick with it:

• Realistic output: The goal isn’t guaranteed historical accuracy, but plausible, pleasing colorizations.
• Fine-grained detail: The product claims it can capture “little wow” elements (small facial features, hands, lighting nuances).
• Customization: A built-in editor lets you adjust the descriptive keywords that guide the result.

In Palette’s own FAQ, it’s explicit about limits: it does not claim to recreate the exact original colors. Instead, it describes the process as
“sophisticated, data-driven guesses”—and then gives you tools to steer those guesses toward your intent.
That framing is important: for archives, genealogy, and restoration work, the best tool is often the one that makes uncertainty visible and editable.

How the text-guided editing works (and how to get better results)

Palette’s customization advice is unusually specific. For stronger colorizations, it recommends describing the image using words associated with objects and
materials rather than naming colors directly.

Example prompts (based on Palette’s guidance):

• Prefer: “a rose dress” over “a red dress”
• Use concrete nouns: grass, ocean, bricks, leather shoes, wooden table
• Include context: “sunny outdoor portrait”, “indoor tungsten lighting”, “old street scene”

It also includes a “Surprise me” option to generate a new description and corresponding look—useful when you want fast creative variety
before you lock in a final aesthetic.

Who it’s for (beyond nostalgia)

Palette is marketed to everyday users reviving family memories, but it also emphasizes professional adoption—particularly documentary and production work.
The site states it’s “trusted in productions” and mentions professional contexts like documentarians and filmmakers.

The founder, Emil Wallner, describes Palette as a project that began while he was at Google and later became his full-time focus, with
ongoing work on increased speed, an API, bulk colorization, and video colorization.

Pricing model: free previews, pay for full resolution

Palette uses a credit system for full-resolution, watermark-free downloads:
1 credit = 1 full-resolution image without a watermark.

What you get for free

• Free trial includes 1 credit (per the pricing page).
• Unlimited color previews
• 21+ color filters
• Customization tools
• But: previews are lower resolution and typically include a watermark

Paid options (as listed on the site)

• Subscription: $72/year (shown as $6/month billed annually) and includes 480 credits/year.
• One-time bundle: $49 for 75 credits, available for 2 years.
• Single-image purchase is described in the FAQ as $1.99 for one image (single purchase rate).

Palette also lists payments via major credit cards, PayPal, Apple Pay, and Google Pay (per the FAQ).

Formats, resolution limits, and browser support (details that matter)

If you’re evaluating Palette for a real workflow—especially bulk restoration—constraints matter. Here’s what the FAQ states:

• Supported formats: PNG and JPEG.
• Max resolution (desktop with HD credits): up to 25 megapixels (example given: 6250 × 4000).
• Max resolution (mobile): up to 9 megapixels (example given: 3000 × 3000).
• Free preview downloads: limited to 0.5 megapixels (example given: 500 × 500).
• Browser notes: Safari for iPhone/iPad; Chrome for Android; Chrome or Safari on desktop. Safari is described as having the highest level of support.

For teams and builders: API + bulk and video options

Palette offers an API and no-code tooling positioned for:

• Bulk colorization (including applying it to Google Drive or Dropbox files).
• App integration (build your own colorization experience).
• Video colorization via a developer Video API (SD, Full-HD, or 4K), and a simpler option via a partner (as described on the API page).

The API documentation notes that the API platform has changed to RapidAPI.

Privacy and data handling: what the site claims

The FAQ states that images are encrypted and that the original and colorized images are permanently deleted once you have received them.
For specifics, Palette directs readers to its privacy policy.

A practical way to test Palette in 10 minutes

1. Upload one portrait and one scene photo (street/interior) to see how it handles skin tones vs. lighting complexity.
2. Run 3–5 different filters to quickly bracket the “mood” you want.
3. Use the text editor to add concrete nouns (materials/objects) and lighting context.
4. Only spend a credit once you’re satisfied with the preview direction—since credits are tied to full-resolution downloads.