FlowRunner
PricingContact
Theme
Start Free

Generate branded images and PDFs from Picnie templates, then compress, resize, crop, convert, watermark, and inspect them. Agents produce and finish visual assets in a single connector.

20 actions API key available
Platform Documentation ↗ Capability data verified 2026-08-01
A client approves a campaign brief and the asset list is ready to produce
Agent calls Get Template to read the editable element names the render has to fill
Agent calls Create Image Collection to render the whole size set from one element map
Agent calls Add Text Watermark so the set is unmistakably a draft before it leaves the agency
Agent calls Get Image Metadata on each supplied source photo and reads the existing EXIF copyright block
Agent posts the watermarked set and the copyright findings to the account channel
One source photo carries a third-party copyright, so the agent asks the account lead before Add Image Metadata stamps the agency as the author

What This Integration Enables

Most rendering services stop at the render. Picnie keeps going, and that is the reason to reach for it. It generates branded images and PDFs from your saved templates, replacing named text and image elements with values from the flow, and it returns the finished URL synchronously so there is no job to poll. Then the same connector compresses, resizes, crops, converts format, applies a filter, removes a background, runs OCR over an image, stamps a text or image watermark, and reads or writes EXIF metadata. A production sequence that would otherwise cross three vendors and two file handoffs runs as one chain of steps inside a flow.

The template model is worth understanding before you build against it. Every generated and edited asset is filed under a project, so Create Project or List Projects comes first. Get Template reports the exact element names for a template, and those names have to match what Create Image sends, so the flow discovers the contract rather than guessing at it. Create Image Collection renders every template in a group in one call, which is how a single design becomes a matched set sized for several networks, though the group id has to be copied from the Picnie dashboard because the service exposes no endpoint that lists groups. This is what Orchestration as a Service looks like at the asset level: the agent runs the whole production line, and hands back the one decision that is about ownership rather than pixels.

Without FlowRunner

Production in pieces Generating the asset, cropping it, compressing it, and branding it happen in four different tools
Drafts that look final A comp reaches a client with no visible mark distinguishing it from an approved asset
Rights taken on trust Nobody checks what the source file already says about who owns it

With FlowRunner

One production line Render, edit, watermark, and stamp all happen inside a single connected sequence
Drafts marked as drafts The watermark is a step in the flow, not something a person remembers to add
Rights checked and escalated Existing copyright metadata is read before anything overwrites it, and a person decides

Use Case Scenarios

A campaign set produced from the record that describes it

Campaign briefs live in Airtable. When a brief moves to approved, the agent reads the headline, offer, and hero image reference, calls Get Template to confirm the element names, and calls Create Image Collection to render the full size set in one pass. Each result is watermarked as a draft, saved to file storage, and written back to the brief record. The account team reviews a complete set against the brief that produced it rather than chasing individual exports.

Preparing customer-supplied photography for publication

A merchant uploads product photography into a listing workflow. The agent calls Upload Asset to bring each file into Picnie, Remove Image Background to isolate the product, Crop Image to the fixed aspect ratio the storefront expects, Resize Image to the catalog dimensions, and Compress Image to keep page weight down. The finished cutout is composited into the branded template with Create Image and pushed to the listing in Shopify. Every photo in the catalog ends up on the same grid because the grid is enforced by the flow, not by whoever uploaded the file.

Reading documents that arrived as pictures

Receipts, signage photos, and screenshots land in a shared folder in Google Drive. The agent calls Extract Text From Image to run OCR over each one, uses the recognised text to classify what the image is and who it belongs to, and routes it accordingly, with Get Image Metadata supplying the capture date and device when the filename says nothing useful. Anything the OCR returns with too little confidence to classify goes to a person with the image attached, instead of being filed somewhere plausible and wrong.

Human-in-Loop Highlight

Add Image Metadata writes EXIF fields into an image, and two of the documented keys are Artist and Copyright. Those are not formatting. They are an assertion about who made an asset and who owns it, written into the file, and once the file is delivered to a client or published to a network, that assertion travels with every copy anyone makes of it. Picnie gives the agent the means to check before it writes: Get Image Metadata reads the full EXIF block of the source, including any copyright the photographer or stock provider already put there.

So the agent reads first. When a campaign asset is built on a supplied photo whose EXIF already names someone else, the agent stops before the stamping step and posts the evidence to the account channel: "Hero image for the Q3 set carries EXIF Copyright: Ferrand Studio 2025, Artist: L. Ferrand. The finishing step would overwrite both with the agency credit. I have not written anything. Confirm the licence covers re-attribution, keep the original credit, or swap the hero for an owned asset?" A person answers, and the answer is recorded against the run. The gate exists because the failure is quiet and durable in the worst combination: the asset looks perfect, the metadata is wrong, and by the time anyone reads the EXIF the file is on a client's server. Agents finish the asset. A human decides whose name goes on it.

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

20 actions

Projects

2
  • Create Project Creates the container that every generated image, PDF, and edited asset is filed under. A project ID is required by all generation and processing operations, so this runs before anything else in a new workflow.
  • List Projects Retrieves every project on the account with its ID, name, type, and creation timestamp. Used to discover the project ID that the generation, editing, watermark, and metadata operations require.

Templates

2
  • List Templates Retrieves the templates saved on the account. Used to discover the template ID and template name that Create Image and Create PDF both require. Returns saved templates only, not the public Picnie catalogue.
  • Get Template Retrieves the full definition of one template including every editable element and its name, with font, colour, alignment, rotation, and position for text elements and the current URL for image elements. Used to build the element list a render will send, so the flow matches the template's real contract.

Image Generation

3
  • Create Image Renders a single image from a template, replacing named text and image elements with supplied values. Rendering is synchronous and the finished URL is returned directly. Element names must match those reported by Get Template.
  • Create Image Collection Renders every template in a template group in one call with the same element replacements applied across all of them, returning an array of URLs. Used to produce a matched set, such as one design sized for several networks. The group ID must be copied from the Picnie dashboard because no endpoint lists groups.
  • Create PDF Renders a template to a PDF instead of an image and returns the finished document URL, taking the same template ID, name, and element replacements as Create Image. Used for certificates, invoices, and reports at the end of an approval flow.

Image Editing

7
  • Compress Image Reduces the file size of an image reachable at a public URL and returns the compressed result. The source format can be preserved or converted to JPG, PNG, or WebP while compressing.
  • Resize Image Resizes an image to exact dimensions or by a percentage of its current size. Aspect ratio can be preserved, and enlargement of images already smaller than the target can be prevented.
  • Convert Image Converts an image into a different file format, with WebP, JPG, and PNG available as targets. Used to normalise mixed uploads into the one format a downstream system accepts.
  • Crop Image Crops an image to exact dimensions with anchor points controlling which part of the source is kept, defaulting to a centre crop. Used to force a fixed aspect ratio across assets that arrived at every shape.
  • Remove Image Background Removes the background from an image and returns a cut-out. Used to isolate products, logos, or people before compositing them into a template with Create Image.
  • Apply Image Filter Applies a single visual filter such as grayscale, negate, emboss, or gaussian blur. Exactly one filter is applied per call, so effects are combined by chaining calls.
  • Extract Text From Image Runs optical character recognition over an image and returns the recognised text as a plain string. Used to read receipts, screenshots, signage, and scanned documents so the flow can classify and route them.

Watermarks

2
  • Add Text Watermark Stamps a line of text over an image, with font file, size, colour, opacity, rotation, padding, and background swatch all specified per call. Used to brand generated assets and to mark drafts and previews before they are shared externally.
  • Add Image Watermark Overlays one image, such as a logo, on another, scaled to fit a maximum width and height and anchored to one of nine positions. Both images must be reachable at public URLs.

Metadata

2
  • Get Image Metadata Reads dimensions, MIME type, colour depth, file size, and the full EXIF block of a JPEG, including camera, GPS, and copyright information. Used to validate uploads and, in the gate above, to find out who a source image already credits.
  • Add Image Metadata Writes EXIF metadata into a JPEG, covering documented keys including ImageDescription, Artist, Software, UserComment, Copyright, Make, Model, and DateTimeOriginal. Used to stamp authorship and copyright onto generated assets, which is why it sits behind a human decision.

Assets

1
  • Upload Asset Uploads an image into Picnie asset storage from a publicly reachable URL such as a FlowRunner file URL, returning a hosted URL usable as a template element image or as the source of any editing operation.

Files

1
  • Save Image To File Storage Downloads a Picnie result image or PDF into FlowRunner file storage and returns both the FlowRunner URL and the original. Used to keep a durable copy inside the flow rather than depending on the Picnie-hosted link.

Frequently Asked Questions

What can FlowRunner do with Picnie?

FlowRunner agents can run Create Project, List Projects, and List Templates in Picnie, plus 17 more actions.

Does connecting Picnie to FlowRunner require OAuth?

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

Can Picnie trigger a FlowRunner workflow automatically?

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

Start building with Picnie

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