Media and transcription
Start with an accessible input file and the correct teamspace. Discover the input fields for the tool you need; some accept URLs, while other operations require additional settings or masks.
Edit an image
Specify which parts must remain unchanged. For mask-based edits, provide the mask in the format required by the current tool schema.
Process a video
Use media_convertVideoFormat for conversion, media_mergeVideos for combining clips, media_removeAudio for removing sound, and media_reverseVideo or media_speedupVideo for playback changes.
media_scriptToVideo and media_textToVideo create videos from textual inputs. media_addBRollsVideo adds supporting footage according to its available inputs. Check the exact tool name in your client, which may normalize casing.
Combine these clips in the supplied order, remove the original audio, and save the result as a new asset. Preserve the source files.
Chain operations using each completed output as the next input. Do not pass an unfinished job ID into a field that expects a media URL.
Transcribe a recording
- Supply an accessible recording to
media_transcribeVideo. - Keep the returned transcription identifier.
- Use
media_getTranscriptionif processing is still underway. - Retrieve the supported download format with
media_downloadTranscriptionFile.
Transcribe this customer interview, then summarize the main objections and include timestamps where available.
Transcription and the assistant’s summary are separate outputs. Keep the original transcript available when quoting the recording.
Assemble a PDF when available
If your connection exposes media_imagesToPdf, use its current schema to assemble images into a PDF. Check availability first; this operation is not present in every connected catalog.
For every operation, preserve job IDs, inspect the result, and save reusable assets.
