Trint
AISend audio and video to Trint for transcription, wait for the transcript, and export it as captions, documents, or spreadsheets. Agents batch-ingest media from URLs with webhook callbacks, file transcripts into shared drives and folders, and machine-translate finished transcripts.
What This Integration Enables
Trint is transcription built for newsroom workloads, and this connector treats it the way a pipeline should: as a step, not an app someone remembers to open. Media enters two ways. Upload Media streams a file synchronously and starts transcribing immediately, the right path for one-off recordings inside a flow. Ingest Media From URL queues asynchronous ingests where Trint pulls the media itself with webhook callbacks reporting progress, the right path for large files and high-volume batches. Because Trint exposes no transcript status field, the connector's Wait For Transcript does what Trint's own example flow does, polling the export endpoint until it stops returning 404, so downstream steps run exactly when the text exists and not before. From there, agents export transcripts as SRT, WebVTT, CSV, EDL, or DOCX, file them into Shared Drives and folders so teams find them without sorting, and localize finished transcripts with machine translation into any language Trint supports.
Without FlowRunner
With FlowRunner
Use Case Scenarios
Interview to Summary Without a Queue
A recording lands in [Dropbox](/integrations/dropbox). The flow gets a temporary link, sends it through Upload Media into the research team's Shared Drive folder, chains Wait For Transcript, and pulls the text with Export Transcript. A summarization step produces decisions and action items, and the result posts to [Slack](/integrations/slack) with a link to the filed transcript. The researcher reads a summary the same day instead of adding the recording to a backlog.
Captions for a Publishing Pipeline
A video pipeline needs broadcast-ready captions, not burned-in guesses. After transcription completes, the agent runs Export Transcript in SRT format, and for localized editions runs Create Translation, waits on the new file id, and exports its SRT as well. The caption files are stored alongside the video in [Google Drive](/integrations/google-drive-api). Every published video carries captions produced from its actual audio, in every target language.
Archive Ingest at Batch Scale
A media library going back years needs to become searchable text. The agent inventories the files, then queues them through Ingest Media From URL so Trint pulls each file directly, tracking progress with Get Ingest Status and filing results by year and program with Create Folder and the folder routing on ingest. Failures surface with their error codes for selective retry, instead of one bad file silently sinking the batch. The batch itself, though, does not start until a person approves it, because ingest is metered spend.
Human-in-Loop Highlight
Transcription is billed: Trint charges transcription minutes against the account that owns the API key, and a batch ingest commits that spend at archive scale. So the archive agent prices the decision before making it. It posts: "The archive sweep found 214 recordings totaling roughly 180 hours of audio. Queueing all of it through Ingest Media From URL bills transcription minutes for every hour against this account. Approve the full batch, start with the 40 files flagged as interview footage, or hold?" The owner decides what the budget buys. The same [human-in-the-loop](/concepts/human-in-the-loop/) review guards Delete Translation, which permanently deletes a translated file; recreating it consumes translation credit a second time.
Agent Capabilities
15 actionsIngest and Transcription
4- Upload Media Streams an audio or video file to Trint from a source URL and starts transcription immediately, returning the new transcript's Trint ID. Supports routing straight into a Shared Drive and folder. Transcription minutes are billed to the account that owns the API key.
- Ingest Media From URL Queues an asynchronous ingest where Trint downloads the media itself, the recommended path for large files and batches. Requires a callback URL that Trint verifies with an ACTIVITY_STARTED event, then posts progress to as the file imports and transcribes.
- Get Ingest Status Reports each ingest's progress through PENDING, IMPORTING, PROCESSING, READY, or FAILED, including error codes on failures and the Trint ID once imported. Covers ingests only, not synchronous uploads.
- Wait For Transcript Blocks until a transcript finishes, determined by polling the export endpoint until it stops returning 404, bounded by a timeout so flows always return. The bridge between upload and export.
Export
2- Export Transcript Exports a finished transcript. Text formats (SRT, WebVTT, CSV, EDL) return as content; DOCX is saved to FlowRunner file storage and returned as a file URL. Fails with 404 while transcription is still running, so pair it with Wait For Transcript.
- Export Story Exports a Trint Story file from Story Builder as Premiere Pro XML, EDL, or DOCX. Story IDs are distinct from transcript IDs.
Library and Organization
5- List Transcripts Lists the files visible to the API key with title, language, duration, excerpt, and metadata, scoped to personal files or to a Shared Drive and folder, with pagination.
- Get Transcript Details Retrieves one file's title, language, duration, excerpt, metadata, and owning Shared Drive and folder. Exposes no progress field, which is what Wait For Transcript is for.
- List Shared Drives Lists the Shared Drives available to the key, the team spaces (formerly Workspaces) transcripts are filed into.
- List Folders Lists folders within a Shared Drive or the key owner's personal files, including parent ids for nesting. Folder ids feed the upload and ingest routing.
- Create Folder Creates a folder, optionally nested, in a Shared Drive or personal files, so archive flows build their own filing structure.
Translation
4- Create Translation Machine-translates a finished transcript into another language as a new file with its own ID, leaving the source untouched. Runs asynchronously; poll the new file until usable.
- Get Translations Lists the translations that exist for a transcript with their file ids and languages. Checked before creating a translation that already exists.
- Delete Translation Permanently deletes a translated file by its own file id. Cannot be undone, and re-creating the translation consumes translation credit again, so it is routed for review.
- List Translation Languages Lists every language Trint can translate into, for validating a target before Create Translation.
Frequently Asked Questions
What can FlowRunner do with Trint?
FlowRunner agents can run Upload Media, Ingest Media From URL, and Get Ingest Status in Trint, plus 12 more actions.
Does connecting Trint to FlowRunner require OAuth?
No. Trint connects to FlowRunner with basic authentication (a username and password), no OAuth flow required.
Can Trint trigger a FlowRunner workflow automatically?
Trint doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with Trint
$100 in credits. No card required. Connect in minutes.