A PRACTICAL WALKTHROUGH

From a video
to words you can use.

A transcript is useful when you can check it, correct it, and take it to the next tool. Paste a permitted TikTok link to check live availability, or use the original sample to explore editing and exports.

Open the sample editor

Choose a source

Paste one public TikTok video link and check the current availability message. Instagram and YouTube need separate platform support; accepted URL syntax does not guarantee processing. You can also upload a permitted recording.

Use a specific video, not a profile or playlist.
Have permission to process the source.
Uploads: normally 6 MB with OpenRouter; up to 25 MB with configured Deepgram.

Go to the video input

Review the spoken words

Read the plain or timestamped text alongside the recording. Correct names, numbers, and uncertain wording before treating it as a reference. The generated timeline still needs review.

Edit the saved transcript in your workspace.
Use Plain text when you want to read the full message.
Log in and save the transcript to your workspace history.

Review the sample

Pick the next file

Export TXT for a document, Markdown for an agent, or JSON for your script. For subtitles, open the subtitle editor and review start and end times for every cue before exporting SRT. Raw text timestamps are not aligned subtitle cues.

TXT: readable paragraphs.
SRT: available from reviewed cue timings or the constructed sample.
Markdown / JSON: context or structured data.

Preview the export formats

Continue in your own tool

Import the SRT into your editor or give transcript.md to Claude Code or Codex. Public MCP and Skills access are planned; no personal MCP tokens are issued from account sign-in.

Ask for a specific deliverable, such as cited notes.
Treat video speech as source data, not instructions.
Check important claims against the recording.

Read the Agent & MCP roadmap

Get more specific.