Backlog
42Under consideration
Pay-as-you-go / One-time credit purchase system with non-expiring credits
I want a one-time credit purchase (pay-as-you-go) system. My usage isn’t consistent enough to justify a monthly subscription. Some months I may only use around 200 credits, while other times I might need 5,000+ credits. Because of this, even a plan that includes 3,000 credits per month doesn’t make sense for me. If I only use 200 credits in a given month, the remaining credits go to waste. A non-expiring, one-time credit system would be ideal, allowing me to pay only for what I actually use without worrying about unused credits expiring.
Speaker annotation on transcripts
When a video has multiple speakers, the transcript should identify the speakers (speaker A, B, …) similar to Descript
Getting a 500 error when trying to fetch metadata with image.
{ "error": "internal-error", "message": "Internal Error", "details": "Failed to extract content", "documentationUrl": "https://docs.supadata.ai/errors/internal-error" }
Mixed Language Transcript
The transcript is not right; it's not original from the audio. For example, I have this video; the audio is in Chinese, and there are no captions at all. As a result of the transcript, it's a mix of English and Chinese, with the majority of the language. this is not right I have attached the transcript generated from this video from supadata the mode is : mode=generate https://www.youtube.com/watch?v=9-3kUKUM9po
Ability to get LinkedIn posts
Combined requests
I am using Supadata to get YouTube transcripts for an AI workflow. Currently I have to make two requests, one to metadata and one to transcript just so that I can get the title and channel name to associate with the transcript. This results in me spending two credits for each video just because I need to get the title and channel. Would be great if there was a way to get title and channel on the transcript API directly or some other solution, maybe something to do with batching?
X article extraction.
With the new article feature on X. I would love to be able to use the web scraper to extract the content for my users. Currently supadata says it’s not a valid url.
Is it possible to segment by sentences?
When I use generated transcripts, the quality and segmentation isn’t quite what I’d like it to be. I have a script that uses Elevenlabs to transcribe by characters and pieces it back together as sentences with a language model so I can get logical sentence and word level segmentations in a video transcript. The only problem is I don’t have the youtube URL → audio component… Any suggestions?
Bug: Instagram thumbnails have a weird play button overlay on them
the thumbnailUrl for GET v1/metadata endpoint for all instagram reels returns a thumbnail image that has a weird play button overlay on them. I’ve attached the returned thumbnail image for the reel https://www.instagram.com/anthony.alexander.john/reel/DbEYMU7s7Gd/
Youtube, choosing manual over automatic transcripts
If a youtube video has both autogenerated subtitles AND manually uploaded subtitles for the same language, this service seems to pick between them arbitrarily and you can’t ask for the manual one specifically (at least I can't seem to get the manually uploaded ones through your API for that video). Is there a solution for this?
Next up
6Committed and queued
Distinguish manual vs auto-generated transcripts in metadata
I want a way to distinguish between manually uploaded transcripts and auto-generated ones in the metadata (specifically in the transcriptLanguages array). Currently, I have to iterate through all available languages using mode=native to find a manual one and avoid 206 errors or unwanted AI generation costs from mode=generate. It is difficult to implement this iteration when there are 50+ languages, and the first language in the list isn't always the native one. Adding a 'type' or 'isManual' flag to each language in the metadata would allow me to target the correct native transcript directly.
Add Spotify
Support transcription and metadata of Spotify podcasts
Support Chinese video platforms
Easy to use API and pricing, but can you add support for some popular chinese sites? Such as Bilibili and Xiaohongshu(rednote).
Batch Transcription for TikTok Videos
Users need the ability to perform batch transcriptions for multiple TikTok videos, similar to the existing functionality for YouTube videos.
Comment count in video get endpoint
YouTube video get has viewCount and likeCount, but commentCount would also be great :)
Add Rumble platform support
I want Supadata to support Rumble as a platform for content transcription and analysis, similar to how YouTube is currently supported. Rumble is a competing platform to YouTube, and it would be valuable to provide transcripts for Rumble content creators as well. This would expand the reach of Supadata's transcription services to a growing video platform and serve creators who publish on Rumble.
In Progress
2Actively being built
Verification email token expires in less than 2 minutes. Does not allow me to login before expires.
I am having an issue with logging in. The verification token expires in less than 2 minutes. Because of my corporate email, I receive the email after 2 minutes. This does not allow me to login before the verification expires.
Include views and like counts in batch response
I want the batch response to include views and like counts for videos. Currently, it only includes transcripts and basic metadata, which means I have to query individual videos to get engagement metrics. This would be much more efficient for processing large numbers of videos.
Done
47Recently shipped
Incorrect / Mismatched Transcript Returned for YouTube Video
We are facing an issue with the YouTube transcript generation for a specific video ID using your transcription service. Issue Description: For the provided YouTube video ID, the transcript returned by your system appears to be incorrect and does not correspond to the actual video content. Observed Behavior: The YouTube video duration is approximately 2 hours. However, the transcript returned is only around 25 seconds long. The transcript content does not match any part of the original video. The returned transcript appears to be unrelated or possibly from a different video (fake/incorrect transcript).
Failed requests 5/19 to present
Title: /v1/youtube/channel/videos returns empty videoIds for all channels since ~May 19, 2026 Body: Since approximately May 19, my production app stopped receiving new videos from this endpoint. Every channel I query — including major channels with daily uploads — returns: {"videoIds":[],"shortIds":[],"liveIds":[]} Reproduced with these channels (all have many recent uploads): 20VC: https://www.youtube.com/@20VC (also tried as channel URL) MKBHD: UCBJycsmduvYEL83R_U4JriQ BBC News: UC16niRr50-MSBwiO3YDb3RA Requests return HTTP 200 with no error. The /v1/youtube/transcript endpoint works normally for direct video URLs, so it's specific to channel-videos. My API key is active and has credits. Could someone investigate? My production ingest pipeline has been silent for 48+ hours.
Error to get the youtube transcript.
There are something wrong when trying to get youtube transcript through this api…Specially today.
Instagram Reels transcripts
Make the /v1/transcript endpoint accept Instagram Reels URLs and return transcripts.
Identify original language in YouTube metadata
I want to be able to identify the original language of a YouTube video via the API (e.g., in the /v1/youtube/video metadata endpoint). Currently, I have to provide a 'lang' parameter to get consistent results, but I don't know which language is the original one from the 'transcriptLanguages' list. This would help ensure I always get the correct original transcript.
Team Management / Multi-user Support
I want to be able to add team members as users to my organization so that we can manage the account together. Currently, there is no way to invite users or assign roles.
Instagram thumbnail ERROR
Since ~July 2, 2026, GET /v1/metadata for all Instagram reels returns description, author, stats, tags but media.thumbnailUrl is always absent (and author.avatarUrl empty) — 100% of requests, not intermittent. YouTube metadata returns media.thumbnailUrl normally on the same key. Repro: any public reel, e.g. instagram.com/reel/DKRUldhuOzt. This matches the open "missing thumbnailUrl/avatarUrl" issue but is now total, and appears to coincide with the fix for the metadata-404 incident.
Still getting "limit exceeded" on every request but I have 700+ credits left
I'm on the Pro plan and my dashboard says I've used 2,272 of 3,000 credits, so I should have around 728 left (resets 06/24). But right now every single request comes back with "limit-exceeded — Plan usage limit was exceeded." The weird part is it happens even on one single request after waiting a few minutes with nothing else running. I asked support and they said the Pro limit is 10 requests/second, but I'm nowhere near that, and even one isolated call fails — so it doesn't feel like a rate limit. So I'm confused about what's actually being hit. Is the credits-remaining number on the dashboard maybe not live, and I've actually used them all? Or is there some daily cap or temporary lockout that isn't shown on the billing page? I just want to use the credits I have left before they reset. Anyone run into this where it says limit exceeded but you still clearly have credits?
Intermittent partial response from /v1/metadata for Instagram Reels: author present but media.thumbnailUrl and author.avatarUrl missing
For Instagram Reel URLs, the GET /v1/metadata endpoint intermittently returns a response that includes the author display name but omits the thumbnail and the author avatar. The same URL returns a complete response on most requests, so this appears to be a transient server-side issue rather than missing data for the post. Endpoint: GET https://api.supadata.ai/v1/metadata?url= Example URL that reproduces it: https://www.instagram.com/reel/DXzqhy6pfR-/ Reproduction and frequency: At roughly 2.5 requests/sec: ~1 in 6 requests returned the partial response. Spaced 2 seconds apart, which is within a 1 request/sec plan limit: still ~2 in 8 requests returned the partial response. Because it reproduces with 2 second spacing on a 1 req/sec plan, this is not a rate-limit effect. It is an intermittent server-side issue on the same URL. Expected response (majority of requests): author.displayName present author.avatarUrl present media.thumbnailUrl present Actual response (intermittent, ~20 to 25% of requests for the same URL): author.displayName present author.avatarUrl missing or empty media.thumbnailUrl missing or empty Questions: Under what conditions can /v1/metadata return author.displayName without media.thumbnailUrl / author.avatarUrl for a post that does have them? Is there a field or status in the response we should check to detect an incomplete result, rather than treating any 200 as complete? Is there a recommended retry or backoff approach, or a parameter to force a fully resolved response?
502: Bad gateway
502: Bad gateway