Pipedream + ReelSDK $0.03 per output minute, first $0.50 free
Render video in Pipedream without hitting the 30 s timeout
ReelSDK renders the video for $0.03 per output minute while the workflow sits suspended, and a suspended workflow costs no credits. Pipedream stops an HTTP-triggered workflow after 30 s by default and 750 s at most, and logs, exports and the event share 6 MB; $.flow.suspend and a POST to resume_url get around both.
- No credits while the workflow is suspended
- Video stays out of the 6 MB step exports
- Plain HTTPS calls that move to any runner
The same job as one call to ReelSDK
export default defineComponent({
async run({ $ }) {
// The workflow pauses at the end of this step until ReelSDK POSTs the finished job
const { resume_url } = $.flow.suspend(15 * 60 * 1000);
const res = await fetch("https://api.reelsdk.com/v1/slideshow", {
method: "POST",
headers: {
Authorization: `Bearer ${process.env.REELSDK_KEY}`,
"Content-Type": "application/json",
Prefer: "respond-async",
},
body: JSON.stringify({
images: [
"https://api.reelsdk.com/f/lib/in/job_x5wepdaqkop4tepq05xp/scene-01.jpg",
"https://api.reelsdk.com/f/lib/in/job_x5wepdaqkop4tepq05xp/scene-02.jpg",
"https://api.reelsdk.com/f/lib/in/fr78av0etm46huw3/river.jpg",
],
size: "1080x1920",
transition: "fade",
voiceover: {
text: "Three photos, one video. This slideshow was rendered by reelsdk and handed back to a Pipedream workflow through its resume URL.",
voice: "luna",
},
captions: "bold",
webhook_url: resume_url,
}),
});
const { job, error } = await res.json();
if (!res.ok) throw new Error(`reelsdk ${res.status}: ${JSON.stringify(error)}`);
return { id: job.id };
},
});
Why ffmpeg breaks here
Keep the video out of the workflow
A Pipedream code step runs inside the workflow's limits: 256 MB of memory by default, 2 GB of disk in /tmp, and 30 s per execution on HTTP and email triggers. You can raise the timeout to 300 s on free tiers and 750 s on paid tiers; past that Pipedream throws a Timeout error. Logs, step exports and the event share 6 MB, so a video returned from a step fails with Function Payload Limit Exceeded, and an HTTP trigger answers 413 Payload Too Large to a body over 512 KB.
$.flow.suspend returns a resume_url and pauses the workflow at the end of the step. Pass that URL to ReelSDK as webhook_url with Prefer: respond-async: ReelSDK answers 202 at once, renders, and POSTs the finished job to the URL, which resumes the workflow. Pipedream doesn't charge for the time a workflow is suspended and bills 1 credit per 30 s of compute at 256 MB per segment, so this workflow uses 2 credits. The render costs $0.03 per output minute plus $0.05 per 1,000 voiceover characters.
Pipedream next to ReelSDK
| Limit | Pipedream | ReelSDK |
|---|---|---|
| Time per execution | 30 s by default on HTTP triggers, 750 s at most on paid tiers | Jobs up to 10 min while the workflow sits suspended |
| Memory and disk | 256 MB by default, up to 10 GB; 2 GB in /tmp | Rendering runs on ReelSDK, not in the step |
| Data between steps | 6 MB for logs, exports and the event | Outputs as URLs, kept 7 days |
| HTTP trigger body | 512 KB by default, then 413 | Inputs by URL, up to 500 MB each |
| Waiting | Suspended up to 24 hours by default, no credits | Webhook: up to 6 attempts over about 7 h, 15 s each |
| Price | 1 credit per 30 s of compute at 256 MB | $0.03 per output minute for slideshows and captions |
Pipedream limits from Pipedream limits, checked November 2026.
Set it up
Two code steps and a key
The workflow needs a trigger and two Node.js code steps. Pipedream code steps only run on Pipedream, so we type-checked both against a declaration of the documented code-step API and ran the same ReelSDK request on the live API.
Add the key
In the workspace settings, add an environment variable named
REELSDK_KEYwith your ReelSDK API key. Every code step reads it asprocess.env.REELSDK_KEY.Environment variable - REELSDK_KEY
- Your ReelSDK API key
Start the render and suspend
Add a Node.js code step named
renderwith the code in the fix block. It suspends with a 15 min timeout, starts the slideshow withwebhook_urlset toresume_url, and returns the job id. 3 photos with a two-sentence voiceover and bold captions come back as a 10-second 1080x1920 video about 16 seconds after the request.Read the finished job
Add a second code step after it. Any POST to
resume_urlresumes the workflow and lands in$resume_data, so this step reads the job back with the key and returns the video and srt URLs, which ReelSDK keeps for 7 days.read/index.mjs export default defineComponent({ async run({ steps }) { // Anyone can POST to a resume URL, so we read the job back with the key const res = await fetch(`https://api.reelsdk.com/v1/jobs/${steps.render.$return_value.id}`, { headers: { Authorization: `Bearer ${process.env.REELSDK_KEY}` }, }); const { job } = await res.json(); if (job.status !== "succeeded") throw new Error(`reelsdk job ${job.id} ${job.status}: ${JSON.stringify(job.error)}`); return { video: job.outputs.video, srt: job.outputs.srt }; }, });
FAQ
Video on Pipedream, answered
Can I run ffmpeg in a Pipedream workflow?
Only inside the step's limits: 256 MB of memory by default, 2 GB of disk in /tmp, and 30 s per execution on HTTP and email triggers, 750 s at most on paid tiers. Logs, step exports and the event share 6 MB, so a video can't pass between steps. Starting a ReelSDK job and suspending keeps the steps to two short HTTP calls.
Why does my Pipedream video step throw Timeout?
HTTP and email workflows stop after 30 s by default and cron workflows after 60 s. You can raise the limit to 300 s on free tiers and 750 s on paid tiers; past it Pipedream throws a Timeout error and stops the workflow. A workflow suspended with $.flow.suspend waits for the webhook instead, up to 24 hours by default.
Does a suspended Pipedream workflow use credits?
No. Pipedream doesn't charge for the time a workflow is suspended. It bills 1 credit per 30 s of compute at 256 MB per workflow segment, so this workflow, one short segment before the pause and one after, uses 2 credits.
Is it safe to give resume_url to ReelSDK?
Any POST to resume_url resumes the workflow, and its body arrives in $resume_data. That is why the read step ignores the body and fetches GET /v1/jobs/{id} with your key before it returns any links.
What happens to Pipedream workflows after March 31, 2027?
Pipedream says Workflows and String shut down on March 31, 2027, and Pipedream Connect is not affected. The ReelSDK calls in these steps are plain HTTPS requests with a webhook, so the same JSON bodies run from n8n, Make, Inngest or your own server.
The same fix on other platforms
All integrationsReady to build?
Ship video from your app today
Your first $0.50 is free, and one key covers every endpoint.
