Browse documentation
Guides
Build search features, research with an agent, or collect video data and media. Start with the result you need.
On this page
What do you want to build or do?
Add search to an application
Return web sources from your backend and expose search as a tool for your agent.
Node.js · HTTP
GET /search + search formResearch a topic with your agent
Collect web and video sources, then give selected evidence to your agent for a cited comparison.
Python · Your agent or model
evidence.json + brief.mdPrepare a YouTube summary with citations
Retrieve existing captions and ask your model for a summary with timestamp links.
Python · Your model
video-evidence/<video>-<language>-<track>.jsonCreate a daily research briefing
Compare sources across runs, then write a briefing with checked citations.
CLI · Python · model
sources.json + briefing.mdDeliver media to your application
Deliver an MP4 to private S3 storage, verify its bytes and give your application a short-lived download link.
Python · CLI · Node.js · S3
Verified file + download linkEnrich a list of YouTube videos
Query distinct videos once, then join metadata and outcomes back to your original application records.
CLI · Node.js
enriched.ndjsonFind and prepare the right data
- Google Search best practices
Refine queries, choose a market and save ranked, deduplicated sources.
- YouTube Data best practices
Select videos, inspect caption languages and preserve timestamped evidence.
- YouTube Download best practices
Choose video, audio or a clip, configure storage and retrieve the delivery record.
Run, save and recover your workflow
- Async & batch requests
Resume background work from saved job IDs and collect every batch outcome.
- Bulk file processing
Process a video ID file and collect each item's result or failure.
- Run recurring agent tasks
Reuse owner access and saved state when your host schedules the next run.
- Save and reuse results
Archive inputs, results, citations and usage for another process to reuse.
- Recover an agent integration
Inspect an interrupted run and recover its outcome before deciding to retry.
- Deploy FlashData workloads
Configure a server or container and verify access and result reuse after a restart.
- Measure API latency and download throughput
Run one measured search, summarize saved samples, or time media delivery to your storage.
yt-dlp Guides and Troubleshooting
How to use yt-dlp
A Linux-first workflow for daily downloads: nightly setup, formats, captions, datasets and resumable collections.
Fix yt-dlp 403 Forbidden
Update nightly, reduce concurrency, check Deno, then test cookies and a healthy exit IP—in that order.
Fix yt-dlp’s “Sign in to confirm you’re not a bot” error
Resolve “Sign in to confirm you’re not a bot” with nightly, Deno, a valid session and a working exit IP.
Fix yt-dlp 429 Too Many Requests
Pause the affected exit IP, reduce extraction and caption request rates, then resume a single worker.
yt-dlp Python API
Runnable Python examples for downloads, metadata, AI datasets and resumable workers.
yt-dlp commands for everyday work
Copy a recipe for video, audio, captions, datasets or collection recovery.
FlashData vs yt-dlp
Compare features, map your existing settings and move a video dataset workflow to managed cloud delivery.