Tinify (TinyPNG)

Compress PNG, JPEG, and WebP images from a public URL, then resize or convert the result to modern formats like WebP and AVIF through the Tinify (TinyPNG) API. Every call reports monthly compression usage so flows stay inside plan limits.

4 actions Basic auth available

Tinify (TinyPNG) website ↗ Platform Documentation ↗ Capability data verified 2026-07-28

A new product with full-resolution photos is published in the storefront

Compress Image shrinks each photo from its public URL and returns the optimized result with its compression ratio

Convert Image produces a WebP variant of each compressed result for modern browsers

Resize Image cuts a Cover-cropped listing thumbnail from the same compressed source

The agent compares before and after byte sizes and reads the monthly compression count from the response

The storefront owner reviews a visual proof sheet before optimized variants replace the live originals

Approved variants upload to the CDN bucket and the product record points at the new URLs

What This Integration Enables

Tinify is the TinyPNG engine as an API: point it at a public image URL and it returns a smaller one, then lets you resize or convert that same compressed result without re-uploading. That shape fits agent pipelines exactly. FlowRunner agents catch images at the moment they enter a system, from a CMS upload, a product feed, or a designer's [Dropbox](/integrations/dropbox) folder, and run them through compress, convert, and resize before anything ships to a page. What makes this connector unusual is that it prices its own honesty into every response: each action returns the account's monthly compression count, so flows always know where the meter stands. - Compress Image handles PNG, JPEG, and WebP sources and reports byte size, dimensions, and compression ratio - Convert Image targets WebP, AVIF, JPEG, PNG, or Smallest (auto), which picks whichever format lands smallest - Resize Image offers Scale, Fit, Cover, and Thumb methods, so one compressed source yields hero, card, and thumbnail - Get Compression Count reads the meter without consuming a compression, the check every batch job runs first - [Human-in-the-loop](/concepts/human-in-the-loop) gates sit where quality and quota decisions belong to a person

Without FlowRunner

Full-resolution images shipped to phones The 8 MB photo from the product shoot is what every mobile visitor downloads

Optimization as a manual chore Someone drags files through a compressor tab when they remember, which is after launch

Plan limits discovered by surprise The monthly compression quota runs out mid-batch and nobody knows which images got done

With FlowRunner

Every image optimized on arrival New uploads are compressed, converted, and resized as a pipeline step, not a favor

One source, every variant A single compressed result fans out to WebP, AVIF, and thumbnail versions in the same flow

Quota as a tracked budget Every response reports the compression count, and batch jobs check the meter before they start

Use Case Scenarios

Product images optimized before the listing goes live

A [Shopify](/integrations/shopify) product is created with photos straight off the camera. The agent runs Compress Image on each public image URL, Convert Image to WebP for storefront delivery, and Resize Image with Cover to cut the collection-grid thumbnail. The variants land in [Amazon S3](/integrations/amazon-s3) and a proof sheet posts to [Slack](/integrations/slack): original size, optimized size, ratio, and side-by-side links. The merchandiser eyeballs the hero shot for artifacts, approves, and the listing publishes with images a phone can load between taps.

The blog pipeline that never ships an 8 MB hero

When a draft is ready, the agent sweeps its image URLs, compresses each one, and calls Convert Image with Smallest (auto) so every asset ships in whichever format is genuinely smallest. Results upload to the CDN bucket via [Amazon S3](/integrations/amazon-s3), the draft's image references swap to the optimized URLs, and the editor gets a summary: nine images, total payload before and after, compressions consumed this month. Page weight stops depending on whether the author remembered to export for web, and the editor's approval is the last human touch the pipeline needs. When a hero image matters enough to warrant it, the same flow flags it for a closer look instead of auto-swapping the reference.

A media library backfill that respects the meter

Years of uploads sit in [Dropbox](/integrations/dropbox) at full resolution. The agent inventories the folder, calls Get Compression Count to see how much of the month's quota remains, and proposes a batch plan in [Slack](/integrations/slack): how many images, how many compressions that costs against the meter, oldest first. The owner approves a tranche, the agent processes exactly that tranche with Compress Image and Convert Image, logs each result to [Google Sheets](/integrations/google-sheets), and stops. Next month, the next tranche. The backfill finishes without ever blowing the budget the live pipeline depends on.

Human-in-Loop Highlight

Every Compress Image call moves the monthly compression count up by one, and the count never moves back down. A three-variant pipeline, compress then convert then resize, spends three units per image, so an enthusiastic backfill across a 4,000-image library is a five-figure draw against a metered plan, committed before anyone sees a result. FlowRunner agents therefore treat the meter as a budget with an owner. Before any batch beyond the routine per-upload flow, the agent calls Get Compression Count, prices the job in compressions, and asks: "1,200 images queued, 3,600 compressions required, 340 used of this month's allowance. Process all, cap at the allowance, or defer?" The owner sets the scope. The same gate guards quality on the other side: an aggressive AVIF conversion that smears a product texture is technically a success and commercially a problem, so hero images get a human eye on the proof sheet before optimized variants replace what customers actually see.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via preferred channel

Human decides

Agent resumes with decision

Agent Capabilities

4 actions

Optimization

Compress Image Compresses a PNG, JPEG, or WebP image from a publicly accessible URL. Returns the optimized image URL with byte size, type, dimensions when available, compression ratio, and the current monthly compression count. Its output URL feeds the resize and convert steps.
Resize Image Resizes a previously compressed image using Scale, Fit, Cover, or Thumb. Scale keeps proportions from one dimension, Fit never enlarges, Cover fills and smart-crops, Thumb pads a standalone thumbnail. Returns the resized URL, final size and dimensions, and the compression count.
Convert Image Converts a previously compressed image to WebP, AVIF, JPEG, PNG, or Smallest (auto), which returns whichever format is smallest. An optional background color fills transparency when targeting formats without it. Returns the converted URL, byte size, content type, and the compression count.

Account

Get Compression Count Reports how many compressions the account has used this calendar month, using a minimal validation request so checking the meter does not spend it. The first call in every batch job on this page.

Frequently Asked Questions

What can FlowRunner do with Tinify (TinyPNG)?

FlowRunner agents can run Compress Image, Resize Image, and Convert Image in Tinify (TinyPNG), plus 1 more action.

Does connecting Tinify (TinyPNG) to FlowRunner require OAuth?

No. Tinify (TinyPNG) connects to FlowRunner with basic authentication (a username and password), no OAuth flow required.

Can Tinify (TinyPNG) trigger a FlowRunner workflow automatically?

Tinify (TinyPNG) doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.

Start building with Tinify (TinyPNG)

$100 in credits. No card required. Connect in minutes.

Start Building Free Talk to us