AutoContent API
AITurn websites, YouTube videos, PDFs, or a bare topic into AI podcasts, videos, infographics, slide decks, and quizzes with the AutoContent API. Agents repurpose one source into several published formats in a single pass.
What This Integration Enables
AutoContent API is a repurposing engine, and Repurpose Request is the action that makes that a real claim rather than a positioning line. Create Content takes a source and produces an artifact: web pages, YouTube videos, PDFs, uploaded files, plain text, a bare topic, saved feed items or prior research results go in, and a podcast, video, infographic, slide deck, quiz, data table or briefing document comes out. Repurpose Request then reuses the sources and settings of a completed request to produce a different artifact, with optional overrides for language, instructions and format-specific options. A finished podcast becomes an infographic, a deck and a quiz without anyone re-collecting a thing.
Underneath that sit the pieces that make the output usable rather than merely produced. Create Podcast From Script renders a two-host episode from dialogue you wrote yourself, with each line prefixed by a speaker marker, which is the path for anything that needs full editorial control. Modify Podcast re-renders existing audio with different voices and optional background music, so an episode gets relocalized or rebranded without regenerating the script. Get Transcript returns speaker-attributed segments for show notes, a search index or accessible captions. Start Deep Research runs a long-form research job whose result ID feeds straight back into Create Content, which is how a briefing gets built from findings rather than from a link someone pasted. Create Project stores a brand description, logo and color palette that every video, infographic, deck and quiz can reuse, so the fan out stays on brand instead of drifting per format.
Without FlowRunner
With FlowRunner
Use Case Scenarios
A research report becomes a content set
A quarterly report is published. The agent calls Create Content with the report URL, targeting a two-host podcast episode. On Content Completed fires with the audio URL. The agent pulls the transcript, checks it against the source, and posts both into the content channel in Slack. Once the editor signs off, a single Repurpose Request per format turns the same approved request into a slide deck for the sales team, an infographic for social, and a quiz for the newsletter. Each finished artifact is written back to the campaign record in Notion. One source, four surfaces, one review.
Private documents that never get a public URL
A team wants to generate content from internal material that cannot be published at a public URL first. The agent calls Upload Source File, which fetches the document from the URL supplied and forwards it as a multipart upload, returning a file ID that Create Content references directly. Before uploading anything it calls List Source Files to recover an existing ID rather than uploading the same document twice. When the engagement closes, Delete Source File removes the uploaded material permanently. Content already generated from the file is unaffected, but any flow still holding that file ID will fail on its next request, so the cleanup step reports which flows referenced it.
Research first, content second
Rather than generating a briefing from whatever a single article says, the agent calls Start Deep Research with a question and, where it helps, project descriptions and attached files for grounding. Results come back as prose or as structured JSON shaped by a format description. The agent polls Get Deep Research until the request completes, then passes the research ID straight into Create Content. The briefing document that comes out is built on findings the team can inspect and re-run, and List Deep Research keeps a record of what has already been investigated so the same question does not get paid for twice.
Human-in-Loop Highlight
The fan out is where this connector needs a person, and it needs one exactly once. Repurpose Request faithfully reuses the sources and settings of a completed request, which is precisely the behavior you want and precisely why an error propagates cleanly. A misread figure in the source report becomes a spoken figure in the episode, a headline number on the infographic, a bullet on slide four, and the correct answer to a quiz question. Four published surfaces, four retraction paths, one unattended call. So the agent produces the first artifact, retrieves the transcript, compares it against the source, and stops before the multiplication: "Q3 report episode is rendered, 18 minutes. Two lines state figures the report does not contain: a 34 percent retention number at 07:12 and a customer count at 11:40. Approve to fan out into deck, infographic and quiz, or send back for a re-render." The editor answers one question and four artifacts follow. The same gate covers voice, because Clone Voice creates a narration voice from a sample and that voice will read whatever the script says, so whose voice is on an episode is a decision a person makes, not a parameter an agent fills in. This is what an automation exception looks like in a content pipeline: the agent did all the work and stopped at the one place where being wrong gets expensive.
Agent Capabilities
24 actionsContent Generation
5- Create Content Starts an asynchronous job that turns source material into a podcast, video, infographic, slide deck, quiz, data table, briefing document or other written output. Sources can be web pages, YouTube videos, PDFs, uploaded file IDs, plain text, a bare topic, saved feed items, prior deep research results, or previously generated podcast episodes. Returns a request ID. Video and slide deck options apply only to their respective output types.
- Get Request Status Reports the progress and results of a generation request. A numeric status of 100 means the job finished and the response carries the hosted artifact URLs plus any generated text. A non-null error timestamp or a non-zero error code is a terminal failure even when the numeric status has not advanced, so check both before retrying.
- List Content Requests Returns the generation requests made by the account with their type, status and title, paginated or returned in full. Used to build a dashboard or reconcile a batch.
- Repurpose Request Reuses the sources and settings of a completed request to produce a different artifact, for example turning a finished podcast into an infographic, slide deck, quiz or data table without re-uploading anything. Optional overrides change the language, instructions and format-specific options. This is the action that turns one approved source into a content set, and the reason approval belongs upstream of it.
- Delete Request Marks a generation request as deleted so it no longer appears in listings. A housekeeping operation for tidying a request history; it does not refund credits already consumed.
Podcasts
7- Create Podcast From Script Renders a two-host podcast from a script you have already written, giving full editorial control over the dialogue. Each line must be prefixed with a speaker marker, which maps to the primary and secondary voices.
- Modify Podcast Re-renders an existing podcast audio file with different host voices and optional background music, which is how an episode gets relocalized or rebranded without regenerating the script. Supply the URL of the audio to modify plus both voice IDs.
- Get Transcript Retrieves the stored transcript of a generated episode as speaker-attributed segments. Used to publish show notes, feed a search index, build accessible captions, or check what an episode actually claimed before it is published.
- List Transcripts Returns every transcript created by the account, optionally narrowed to those queued, failed, succeeded or processed, paginated or returned in full.
- List Voices Returns every narration voice available to the account, covering stock voices and any cloned on a PRO plan. Each entry exposes the numeric ID that the content, script and modify actions expect. Pass the numeric ID, never the voice name.
- Clone Voice Creates a custom narration voice from an MP3 sample hosted at a public URL, so a brand or presenter voice can be reused across every episode the account generates. A PRO feature that runs asynchronously. Samples are limited to 10 MB.
- Remove Cloned Voice Deletes a cloned voice owned by the account and frees the slot it occupied. Only voices you created can be removed; stock voices are permanent. Any flow still referencing the removed voice ID will fail on its next generation call, so audit before removing.
Research
3- Start Deep Research Kicks off a long-form research job on a question, optionally grounded in project descriptions and attached files. Results come back as prose or as structured JSON shaped by a format description, and the resulting research ID feeds straight into Create Content to build a podcast, deck or infographic from the findings.
- Get Deep Research Retrieves a single deep research request and its findings. Deep research uses its own status endpoint rather than the content status endpoint, and a non-zero error code indicates a terminal failure.
- List Deep Research Returns every deep research request made by the account with its prompt, status and completion timestamps. Used to pick up a research ID for a follow-up generation, or to audit what has already been investigated.
Video
2- Create Video Short Generates a short-form talking-avatar video from supplied text or from saved feed items, suitable for social channels. One avatar is required and a second can be added for a two-presenter format; subtitles are burned in by default.
- List Avatars Returns the presenter avatars available for video generation, each with the identifier that Create Video Short expects. Call this before building a video flow so avatar selection is grounded in what the account can actually use.
Sources and Projects
5- Upload Source File Uploads a private source document so it can be referenced by ID instead of being published at a public URL. The file is fetched from the URL you supply and forwarded as a multipart upload, and the returned file ID can be used as a resource in Create Content or attached to a deep research request. Uploads may be up to 200 MB.
- List Source Files Returns the private files already uploaded under the authenticated token, so a flow can recover an existing file ID rather than uploading the same document twice.
- Delete Source File Permanently removes an uploaded source file from AutoContent storage. Content already generated from the file is unaffected, but any flow that still references the file ID will fail on its next request.
- List Projects Returns the projects owned by the account, newest first. A project stores a brand description, logo and colour palette reusable as context and branding across videos, infographics, slide decks and quizzes.
- Create Project Creates a reusable brand project holding a description, logo and colour palette. Supplying a website URL without a description makes AutoContent fetch the site and write the description for you. Reference the returned project ID from Create Content or Start Deep Research to keep branding consistent across every format in a fan out.
Account
2- Get Usage Reports the account's total credit allowance, credits already consumed, the daily podcast limit and how many podcasts have been created today. Used to guard a batch job before it starts, or to alert when the remaining balance drops below a threshold.
- List Webhooks Returns the HTTPS webhook endpoints currently registered on the account, each with its ID and URL. Used to audit which callbacks are live, including those the trigger created automatically.
Triggers
1 triggersEvent Triggers
1- On Content Completed Fires when a generation request reaches a terminal state, delivering the finished artifact URLs without any polling. AutoContent sends callbacks for both successes and failures, so use the Successful Only toggle to ignore failed jobs and the Callback Data filter to react only to jobs your flow tagged when it submitted them. The extension registers an HTTPS webhook on your account when the flow is deployed and removes it when the flow is undeployed.
Frequently Asked Questions
What can FlowRunner do with AutoContent API?
FlowRunner agents can run Create Content, Get Request Status, and List Content Requests in AutoContent API, plus 21 more actions.
Does connecting AutoContent API to FlowRunner require OAuth?
No. AutoContent API connects to FlowRunner with an API key, no OAuth flow required.
Can AutoContent API trigger a FlowRunner workflow automatically?
Yes. AutoContent API supports 1 trigger that can start a FlowRunner workflow automatically.
Start building with AutoContent API
$100 in credits. No card required. Connect in minutes.