Bannerbear
Developer ToolsGenerate branded images, videos, collections, animated GIFs, and website screenshots from Bannerbear templates. Renders run asynchronously against the Bannerbear API v2.
What This Integration Enables
Agents can produce on-brand creative as a byproduct of the workflow that needs it, rather than as a separate design request. Create Image renders a template with per-layer modifications and can wait inline for completion, Get Image polls for the finished asset, and List Templates plus Get Template reveal the editable layer names so an agent knows exactly what it can change. Create Video, Create Collection, Create Animated GIF, and Create Screenshot extend the same template-plus-modifications model to other formats, and Get Account verifies the key and remaining quota before a batch. Because every render is driven by an approved template, the connector generates volume without generating off-brand output, and the finished asset waits for a human before it is used.
Without FlowRunner
With FlowRunner
Use Case Scenarios
Promotional graphics from product data
A new product is marked ready in the catalog. The agent reads the fields that map to template layers, calls Create Image with Wait For Completion so it returns the finished asset inline, and confirms the image URL is populated. It posts the rendered graphic into the marketing review channel. A brand owner approves, and the downstream workflow uses the approved asset. The graphic that used to be a design ticket is now a rendered file waiting on one yes.
Batch creative from a template set
A campaign needs the same message rendered across a set of formats. The agent calls Create Collection against a template set with shared modifications, polls Get Collection for the per-template image URLs, and assembles the full set for review. The team approves the batch as a set rather than commissioning each size individually.
Website preview screenshots for a report
A monitoring workflow needs a visual of a set of public pages. The agent calls Create Screenshot for each URL and collects the results when the render webhook fires. The screenshots feed a report that a person reviews before it is circulated, so the automated capture never becomes an unreviewed claim.
Human-in-Loop Highlight
A rendered asset is one publish away from a customer's screen, so the connector treats the generated creative as a draft, not a deliverable. This is human-in-the-loop: an execution pattern where AI agents pause autonomously, assemble the relevant context and the decision choices available, route to a human via their preferred channel, and resume the moment the human responds. When Create Image completes, the agent posts the finished asset into the review channel and asks: "Rendered from [template] for [product]. Approve for publish, or send back?" The brand owner approves or rejects, and only an approved asset moves downstream. The agent renders freely; it never ships. Rendering the image is the mechanical part. Deciding that a specific generated graphic represents the brand well enough to put in front of a customer is the judgment part, and it belongs to a person. Placing that approval gate between generation and publish, consistently, across every asset an agent produces, is an orchestration problem. Orchestration as a service is the category built to sit above the render step and the publish step and hold the human decision between them, and FlowRunner is built for that layer.
Agent Capabilities
12 actionsImages
3- Create Image Renders an image from a template by supplying modifications for named layers. Renders asynchronously and returns a pending object with a uid; set Wait For Completion to poll inline, poll Get Image, or supply a webhook URL. The primary render action and the asset a human approves.
- Get Image Retrieves a single rendered image by its uid, including status and the image URL once completed. Used to poll for completion after Create Image.
- List Images Lists images previously rendered in the current project, most recent first, with pagination. Used to audit or reuse prior renders.
Templates
2- List Templates Lists templates available in the current project, optionally filtered by name or tag, each exposing its uid and editable layer names. Used to pick a template for a render.
- Get Template Retrieves a single template by its uid, including the available_modifications array that lists every editable layer name. Used to know exactly which layers a render can change.
Videos
2- Create Video Renders a video from a video template using an input media URL and per-frame modifications. Renders asynchronously; poll Get Video or supply a webhook URL.
- Get Video Retrieves a single video by its uid, including render progress and the video URL once completed. Used to poll for completion after Create Video.
Collections
2- Create Collection Renders a set of images in one call from a template set, applying the same modifications across every template. Renders asynchronously; poll Get Collection or supply a webhook URL.
- Get Collection Retrieves a single collection by its uid, including the per-template image URLs once completed. Used to poll for completion after Create Collection.
Screenshots
1- Create Screenshot Captures a screenshot of a public website URL. Renders asynchronously; supply a webhook URL to be notified when the screenshot completes.
Animated GIFs
1- Create Animated GIF Renders an animated GIF from a template by supplying up to 30 frames of layer modifications. Renders asynchronously; poll or supply a webhook URL.
Account
1- Get Account Retrieves the current project account details, including quota usage and plan. Used as a connection check to verify the API key is valid before a batch.
Start building with Bannerbear
$100 in credits. No card required. Connect in minutes.