Pictory
AITurn scripts and articles into finished videos with Pictory. Agents build storyboards from text, render them into captioned, narrated MP4s, and poll job status until the downloadable video URL is ready.
What This Integration Enables
Most content teams are sitting on their best video material already; it is just formatted as text. Pictory's pipeline turns that archive into finished video without an editing suite: Create Storyboard breaks a script into scenes and matches visuals, captions, and voice-over automatically, and Render Video produces the final MP4 with a downloadable URL. FlowRunner runs that pipeline as a governed production line, and the pipeline's two-step design is the point: a storyboard is inspectable before the render happens, which is exactly where a person belongs. The economics follow the same split. Storyboarding is the cheap, reversible half; rendering is the paid, public half. A pipeline that treats those as one step optimizes for volume and ships its mistakes. A pipeline that pauses between them gets the volume and keeps the judgment. - Turn blog posts, articles, and AI-drafted scripts into captioned, narrated videos inside a content pipeline - Feed scripts drafted by upstream workflow steps, so the storyboard starts from copy written for the ear - Build storyboards from structured scene texts when you want control over the beats - Batch-produce localized variants with captions and voice-over across supported languages - Poll jobs to completion and hand finished video URLs to publishing and notification steps - Keep every render's job ID, source, and output URL logged for the content calendar
Without FlowRunner
With FlowRunner
Use Case Scenarios
The blog-to-video line
When a post goes live on [WordPress](/integrations/wordpress), the flow pulls the article, feeds it to Text to Video, and polls Get Job Status until the storyboard is ready. The editor gets the scene breakdown for a two-minute review. After approval, Render Video produces the MP4, the flow uploads it back through the site's media library, and the post gets its video embed. The article that published Tuesday morning has its video version Tuesday afternoon. A failed render retries once, then files the job ID and error for a person instead of letting the article quietly stay video-less.
One announcement, every channel's format
A product announcement script is written once. The flow runs Create Storyboard with structured scene texts so the beats land where the team wants them, renders the approved cut with Render Video, and hands the videoURL to [YouTube](/integrations/youtube) for the channel and to [Slack](/integrations/slack) for the launch room. Job IDs and output URLs log to [Google Sheets](/integrations/google-sheets), so the calendar shows exactly which asset came from which script revision.
The evergreen library, refreshed on schedule
A quarterly flow walks the top-performing help articles, storyboards each with Text to Video, and queues the storyboards for editorial review in batches. Approved storyboards render overnight, with Get Job Status collecting each videoURL as it completes and filing failures for retry instead of silently dropping them. The support site's video coverage grows every quarter without a single editing session being scheduled.
Human-in-Loop Highlight
The step that costs real money and real brand equity is Render Video. Storyboarding is cheap to redo; a render consumes paid capacity, and its output is a finished asset with a voice-over and stock visuals that will carry the company's name to YouTube an hour later. The specific failure is well known to anyone who has automated video: scene-matching reads a metaphor literally, and the article about "killing your onboarding friction" gets storyboarded with imagery nobody wants attached to the brand. So the FlowRunner pipeline holds every storyboard at the review gate. The editor sees the scene list with its selected visuals, swaps the misreads, and approves; only then does Render Video spend the render, and only rendered assets a person approved reach the publish step. The machine does the storyboarding, polling, rendering, and filing. The judgment about what the brand looks like on screen stays human.
Agent Capabilities
4 actionsStoryboarding
2- Create Storyboard Creates a storyboard from a script or plain text, breaking it into scenes with automatically matched visuals, captions, and voice-over. Accepts free-form scene texts or a single text block, and returns a jobId to poll. The asynchronous first half of every render pipeline.
- Text to Video Convenience action that storyboards a single block of raw text or a script directly, with Pictory splitting scenes and selecting visuals automatically. Returns a storyboard jobId to poll before rendering. The fastest path from article to reviewable storyboard.
Rendering
1- Render Video Renders a completed storyboard into the finished video, returning a render jobId to poll. When the render completes, the job data contains the downloadable video URL. The paid, brand-facing step this page's human gate protects.
Job Management
1- Get Job Status Retrieves the status of any asynchronous Pictory job by jobId, for both storyboard and render jobs. Poll until a terminal status; completed render jobs carry the videoURL in their result data. The step that turns Pictory's async pipeline into a reliable production line.
Frequently Asked Questions
What can FlowRunner do with Pictory?
FlowRunner agents can run Create Storyboard, Render Video, and Get Job Status in Pictory, plus 1 more action.
Does connecting Pictory to FlowRunner require OAuth?
Yes. Pictory connects to FlowRunner with OAuth 2.0, so agents authenticate without handling raw credentials.
Can Pictory trigger a FlowRunner workflow automatically?
Pictory doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with Pictory
$100 in credits. No card required. Connect in minutes.