SCRNIFY

SCRNIFY is a pay-as-you-go screenshot and video capture API. Workflows render any public URL into PNG, JPEG, or WebP images or MP4, WebM, and GIF recordings, and generate signed capture URLs safe to publish.

Verified 4 actions API key available

SCRNIFY website Platform Documentation Capability data verified 2026-08-25

A release note, changelog entry, or customer-facing update is ready to publish

Capture Screenshot renders the element the message is actually about, using a selector rather than the whole viewport

Agent confirms the selector resolved and the capture shows the state the copy claims

Get Capture URL builds the signed URL that will render inside the published page

Agent reports the cache window that URL was signed with, and therefore how long it serves the same frame

The draft goes to its owner with the capture and the cache window shown together

The owner sets the window against how long the message will be read, then approves the publish

What This Integration Enables

SCRNIFY renders a public URL into an image or a video, and two things about it shape how it should be used in a workflow. The first is that video is billed by requested duration in seconds against the same pool a screenshot draws one unit from, so a recording is not a slightly more expensive screenshot. The second is that cache hits and failures cost nothing, which turns Cache TTL from a performance setting into the primary cost control.

The operation the product is really built around is Get Capture URL. With an API Secret configured, it returns a URL signed with HMAC-SHA256 over the exact query string, which means a recipient can load it and cannot change what it captures. That is the difference between a live product image you can put in a published page and a link that lets anyone spend your account. FlowRunner agents do the capture, the selector work, and the packaging. The decision about what gets published with a live capture inside it, and for how long that capture stays frozen, is where human-in-the-loop belongs.

Without FlowRunner

Stale attachments A screenshot pasted into a page shows the product as it looked the week the page was written

Capture cost nobody modeled A recording loop bills by the second and the bill arrives before anyone notices the pattern

Publishable URLs that are credentials The quickest way to embed a live capture also hands the reader a key that captures anything

With FlowRunner

Captures that refresh on load A signed URL renders the current page when the reader opens it, within the window it was signed for

Cost as a set parameter Cache windows and recording durations are chosen on the step, not discovered on the invoice

Signed and scoped A signed URL works only for the parameters it was signed with, so a recipient cannot retarget it

Use Case Scenarios

A live product image inside a published page

A marketing page hosted on Webflow shows the current dashboard rather than a screenshot somebody exported in March. The agent builds the embed with Get Capture URL against a named element selector, which captures one panel rather than the whole viewport and waits for that element to become visible before firing. That last behavior is worth more than any timing option: the selector is a wait condition as well as a crop. With an API Secret set, the URL is signed, so the page can be public and the account cannot be spent by its readers.
A recording that shows what the bug actually did

A failure lands in Sentry and the agent reproduces the path in a capture. Capture Video records the sequence as an MP4 and Save Capture To FlowRunner File pulls the bytes into workspace storage rather than leaving them on a public asset host. The recording attaches to a Linear issue and posts to Slack. The duration is set deliberately, because a sixty second recording costs sixty times a screenshot from the same pool, and most reproductions are done in eight.
The same capture, taken once

An internal dashboard image appears in a daily digest, a status page, and a Slack unfurl. Without a cache window that is three captures a day forever. With one, the first render pays and the rest are cache hits, which the vendor states cost nothing. The same setting is what makes a video embed affordable at all. The tradeoff is not free, and the next scenario is where it gets decided.

Human-in-Loop Highlight

Signing makes Get Capture URL safe to publish, and it also makes Cache TTL mandatory, which is where the interesting consequence hides. A signed capture URL serves the frame it cached for the whole window, up to thirty days. Publish one on a status page or in a customer email with a long window and you have not embedded a live view of the product, you have embedded a photograph of it that will keep insisting the product looks that way after it stops. The savings and the staleness are the same setting. So the agent builds the capture and stops: "Changelog entry for the new billing screen is ready. The embedded capture is signed with a 30 day cache window, so it will serve today's frame until October 4 even if the screen changes. This page is linked from the pricing nav. Shorten to 1 hour, attach a workspace file instead, or publish as built?" The person who owns the message knows how long people will be reading it. The agent does not, and a published page is not something you get to take back.

Agent processes routinely

Detects exception requiring judgment

Clear match Continues automatically

Ambiguous Routes to human via email

Human decides

Agent resumes with decision

Agent Capabilities

4 actions

Captures

Capture Screenshot Renders a URL into a PNG, JPEG, or WebP image and returns its public asset URL. Height is required unless full-page capture is on, and a Selector captures one element rather than the viewport while waiting for that element to become visible, which is a more reliable way to wait for content than any timing option.
Capture Video Records a URL as an MP4, WebM, or animated GIF and returns its public asset URL. A video costs its requested duration in seconds rather than a single unit, drawn from the same pool as screenshots, so the duration field is a spend decision. The maximum is sixty seconds.
Save Capture To FlowRunner File Renders a URL and stores the resulting image or video in FlowRunner's own file storage. The form to use for anything confidential, because SCRNIFY's asset URLs are public and a capture of an internal admin screen is readable by anybody who learns the link.
Get Capture URL Builds a ready-made SCRNIFY capture URL, signed when an API Secret is configured, without calling the API or spending a usage unit. Signing is what makes the result publishable: a signed URL works only for the parameters it was signed with, so a recipient cannot retarget it. Without a secret, the URL carries the bare API key and is a spendable credential.

Frequently Asked Questions

What can FlowRunner do with SCRNIFY?

FlowRunner agents can run Capture Screenshot, Capture Video, and Save Capture To FlowRunner File in SCRNIFY, plus 1 more action.

Does connecting SCRNIFY to FlowRunner require OAuth?

No. SCRNIFY connects to FlowRunner with an API key, no OAuth flow required.

Can SCRNIFY trigger a FlowRunner workflow automatically?

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

Start building with SCRNIFY

Free plan, no card required. Connect in minutes.

Start Building Free Talk to us