Generate images and videos from Bannerbear templates: render, poll, batch, and manage templates, webhooks, assets, and workflows.
Bannerbear ships in the w6w first-party pack. It declares 61 actions, 2 health checks, and the host runs its code in a sandbox that never sees the credential.
io.w6w.bannerbearBannerbear turns a design (an Image Template or Animation Template built in its dashboard editor) into a rendering API: send template modifications — swap in text, images, colors — and get back a finished JPG/PNG/PDF or MP4/MOV. This app covers the full v5 REST surface: create and poll Images and Animations, submit up to 100 images at once as a Batch, manage Image and Animation Templates and their versions, install Publications from the public template library, mint self-service Instant URLs that render on GET with no API call, register Webhooks instead of polling, and run the 16 standalone video/image Tools (background removal, subtitles, trims, overlays, concatenation, AI image generation, and more) plus multi-step Workflows that chain several of those together server-side.
Two hosts matter: api.bannerbear.com renders asynchronously (submit, then poll or subscribe a webhook) everywhere, while sync.api.bannerbear.com accepts only image creation and returns the finished render inline — useful for a workflow step that needs the URL immediately and can tolerate a 408 on a slow render. Every action here defaults to the async host; nothing silently blocks a workflow run waiting on a video.
Health is reported from what the API itself exposes: /account returns live plan quota (current vs. max renders this billing period) with no scope requirement, so it doubles as this app’s credential probe and its quota check. Bannerbear’s own status page (status.bannerbear.com) is a stale client-rendered app with no reachable JSON, RSS, or Atom feed, so vendor-status is declared unavailable rather than faked.
Three routes to the same 61 actions. The Workflow tab is generated from Bannerbear's own manifest and carries its real ids, so it is copy-pasteable; the Code and CLI examples are the same call for any action on any app, so every app-specific value in them is a blank you fill in.
animation-create Render a video from an Animation Template. Not idempotent — every call starts a new render.
animation-template-create Create an Animation Template's metadata and canvas size. Its layers/timeline can only be authored in the Bannerbear dashboard — not idempotent, every call creates a new template.
animation-template-delete Delete an Animation Template. Idempotent in effect: a repeat call 404s harmlessly.
animation-template-get Get an Animation Template, including its full canvas config — the layer names/ids to target from animation-create's modifications.
animation-template-update Update an Animation Template's name, description, tags, size, or frame rate.
asset-check Check up to 100 SHA-256 content hashes against already-uploaded assets, to skip re-uploading files the workspace already has.
asset-upload Upload an image, video, audio, or PDF file (as base64) to use as a source in a template modification. Idempotent: Bannerbear matches by content hash, so uploading the same bytes twice returns the existing asset rather than creating a duplicate.
batch-create Render up to 100 images in one call. Not idempotent — every call creates a new batch.
image-create Render an image from an Image Template. Not idempotent by default — pass the same workflow invocation id as metadata to correlate retries yourself.
image-template-create Create an Image Template. Not idempotent — every call creates a new template.
image-template-delete Delete an Image Template. Idempotent in effect: a repeat call 404s harmlessly.
image-template-get Get an Image Template, including its full canvas config — the layer names/ids to target from image-create's modifications.
image-template-update Update an Image Template's metadata and/or canvas config. Idempotent: sending the same fields twice leaves the same result.
instant-url-create Create a self-service render URL bound to a template. Not idempotent — every call creates a new URL, and its signing_key (when security is signed) is returned ONLY here.
instant-url-delete Delete an Instant URL. Idempotent in effect: a repeat call 404s harmlessly.
instant-url-update Update an Instant URL's binding, mode, security, status, scale, rate limiting, version pin, render cap, or expiry. name and template are required even when unchanged.
publication-install Install a publication into this workspace as a new Image Template. Not idempotent — each call installs another copy. Trial accounts are capped at 3 templates.
tool-concat-videos Join two or more videos end to end, with an optional transition.
tool-create-gif-preview Make an animated GIF preview of the start of a video.
tool-create-pdf Combine JPG, PNG, or PDF pages into a single PDF, in the order given.
tool-create-video-slideshow Turn two or more images into a video slideshow with transitions.
tool-generate-ai-image Generate an image from a text prompt. Credit cost varies by model.
tool-generate-voiceover Text-to-speech, up to 2000 characters, via a pre-made ElevenLabs voice.
tool-remove-bg Cut the subject out of a PNG/JPG image against a transparent background.
webhook-delete Delete a Webhook. Idempotent in effect: a repeat call 404s harmlessly.
webhook-update Update a Webhook. Bannerbear requires name and url even when unchanged — this replaces the record, it does not merge partial fields.
workflow-run-create Start a Workflow with the inputs it declares (see workflow-get). Not idempotent — every call starts a new run. To process many rows, call this once per row.
A workflow step names the app and the action, and the editor fills in the
connection when you pick one. This is the Step shape from the
workflow spec, carrying Bannerbear's real ids.
{
"manifestVersion": "2",
"name": "bannerbear-example",
"steps": [
{
"id": "animation-create",
"uses": {
"app": "io.w6w.bannerbear",
"action": "animation-create",
"connection": "conn_YOUR_CONNECTION_ID"
},
"with": {
"template": "<template>"
}
}
]
}animation-create account-get animation-get animation-list animation-template-create +56 more actions available
Every app-specific value here is a blank you have to fill in. An
app action is reached through the connection that authenticates it, so the
address is a connection id, not the app id — and connections belong to your account,
so a public page cannot know yours. Create one for Bannerbear, then fill in
the three blanks: conn_YOUR_CONNECTION_ID, the action key, and the
parameters that action declares. The call itself is real — the shape is transcribed
from the studio's own snippet builder, which prints the same kind of blanks — but
nothing in it is specific to Bannerbear. The Workflow tab is where this app's
real ids are.
npm install @w6w/sdkyarn add @w6w/sdkpnpm add @w6w/sdkdeno add npm:@w6w/sdkimport { W6wClient, isActionRun } from "@w6w/sdk";
// Reads W6W_BASE_URL and W6W_TOKEN from the environment when omitted.
const client = new W6wClient();
const envelope = await client.run({
urn: "conn_YOUR_CONNECTION_ID",
action: "animation-create",
payload: {
template: "<value>",
// modifications: "<value>",
// formats: "<value>",
// metadata: "<value>",
},
});
if (isActionRun(envelope)) console.log(envelope.value); npm install -g @w6w/cli w6w run conn_YOUR_CONNECTION_ID --action animation-create --payload '{"template":"<value>"}' Give an AI agent Bannerbear — without giving it Bannerbear's credentials. One MCP endpoint exposes every app, function and workflow the caller is entitled to, as tools it can discover and run. Access is granted per team while we onboard.
One tool call{
"name": "w6w_invoke",
"arguments": {
"ref": "app:io.w6w.bannerbear#animation-create",
"input": {
"template": "<template>"
}
}
}
Every tool names its target with a single ref. The
app: form above doesn't name a connection at all — the
host resolves which of the caller's Bannerbear connections to sign
with, and refuses rather than guesses when the answer is ambiguous.
The token is attached host-side, at the moment of the call. It is never a tool argument, never in the model's context, and never in a transcript — so a prompt injection has nothing to exfiltrate.
Tools are derived per end user from what that person has actually connected and is entitled to — not one shared bot identity carrying the union of everyone's access.
Multi-step work runs on the workflow engine and returns a run handle the agent can poll — retries, branching and state survive the conversation that started them.
Bannerbear's declared health checks are on the surface too, so an agent can tell "the vendor is down" from "your credential expired" before it burns a retry on either.
The MCP surface is part of the hosted platform. Bannerbear itself is MIT, and the runtime that executes it is source-available (FSL).
Bannerbear declares its own checks, so its health is a property of the app rather than something the host guesses at.
status.bannerbear.com is a stale client-rendered status page (last-modified 2023) with no reachable JSON, RSS, or Atom feed — every candidate path answers the same SPA shell.
Current vs. max renders this billing period, read from GET /account.