Feature-requests

Full-length video access via API — /download returns only ~34 min
Hi tl·dv team, I'm writing from SQUAADS, a digital agency using the tl·dv API ( pasta.tldv.io/v1alpha1 ) to build an automated video-clipping workflow on top of our recorded meetings. First of all — great product, the transcript + highlights endpoints work flawlessly. The issue: GET /meetings/{id}/download returns only the first ~34 minutes (~2050 seconds) of each recording. Most of our meetings run 1–2 hours (e.g. our monthly training sessions), so we're currently unable to reach content beyond minute 34 — which is exactly where the most interesting parts often are. A few observations that might help you point us in the right direction: The endpoint redirects (302) to a signed URL on puttanesca-v0.tldv.io/v1/download.mp4?t= ... serving an fMP4 stream. The signed JWT references a master.m3u8 on media-files.tldv.io , which suggests the source recordings are stored as HLS. The CDN doesn't seem to honor byte-range requests, so partial streaming isn't an option on our side. What we'd like: 1 - Is there a way to download the full-length recording? Any query parameters (e.g. ?start=/?duration=), a different endpoint, or direct access to the HLS segments would be perfect. 2 - Alternatively — do you have (or plan) a server-side clip generation API? (e.g. create a clip by start/end time and get back an MP4 URL) A bit of context on how we use it: we've built an AI pipeline that searches our meeting transcripts, detects the exact timestamps of a topic, cuts the clip locally with ffmpeg, generates a custom thumbnail with GPT-image-2, and publishes it as an unlisted YouTube video — fully automated. Full-length access would let us clip anything from any meeting, not just the first half hour. Happy to share more details about our use case if useful. Thanks for the great tooling! Best regards, Yunior González (well... this time, hermes ;) ) CTO, SQUAADS — https://squaads.com/
0
Load More