Getting started with plycdn
plycdn hosts your video, audio and documents, delivers them fast, and knows what is inside them. Everything you store can be searched by meaning and by wording, in many languages including Indian languages, and a written answer cites the exact moment in a recording or the page of a document it came from.
This page is the shortest road to a first search. Everything else is linked from the end of it.
1What you can do with plycdn
- Store and deliver files. Storage buckets hold whatever you put in them (video, audio, documents, anything
else) and serve it back publicly or through a signed link (plycdn links), at a
plycdn.nethostname or on your own domain (cdn.example.com): see Your own domain in the Storage guide. - Use the tools you already have. Buckets can also be reached with the S3 and Azure tools you use today: the
AWS CLI, boto3 and the other AWS SDKs, rclone, s3cmd, Cyberduck, the Azure CLI and the Azure Storage SDKs, at hosts such as
s3.{label}.plycdn.netand{account}.blob.{label}.plycdn.net. Those tools sign with a storage credential, created in the dashboard under Developers → Storage credentials, never with your API key. The compatibility guide lists the hosts, what is supported, the limits and how it is charged, with an rclone recipe for moving an existing library in and notes on request signing. - Stream video. Upload a recording once and plycdn encodes it at several qualities (never above the
original), makes a poster and a scrubbing strip, and serves it from a
plycdn.netaddress through links only your backend can sign. A video plays as soon as its lowest quality is encoded, and at every quality once it isready. Your pages show it with the packages' player: keyboard-operable, readable by screen readers, remembering where each viewer stopped, and renewing its links so a four-hour lecture plays through. Your backend hears when each video is playable, ready or failed through signed webhooks, instead of polling. Encoding is counted in minutes of source video, and a video that fails is never counted. Videos live in video libraries, a kind of storage bucket; the Video guide covers the whole path. Courses that must be watched in order: Serial playback, and step by step. Each viewer's name, email or your ID for them drawn over the video, so a recording shows who it came from, and your logo in a corner: Viewer watermark and your logo, in every plan. - Go live. Broadcast from OBS or any encoder, or straight from a web page, and viewers watch a few seconds behind, or in under a second for an interactive audience. Each broadcast becomes an ordinary video when it ends, and can be made searchable. A second encoder as a backup, a rewind window of up to two hours, and viewer and health figures are part of it; live captions, restreaming to other platforms, ad insertion and content protection come with the plans that include them (Billing & usage → Usage → Plan shows yours): Live streaming, and step by step. In the dashboard it has its own place: Live streaming → Live streams.
- Keep your logs, metrics and traces. Send them from your applications with OpenTelemetry, Prometheus remote write or plain JSON lines, keep them for 7 to 365 days in the region you choose, and read them back with SQL or KQL, an export, a live tail of your logs and a view of any one trace, with an optional nightly copy into a bucket of your own. It is part of every plan, and your first log line can be in place in about five minutes: Observability quick start.
- Search what is inside. Content Search indexes recordings (what is said, and text shown on screen) and documents, and finds passages by meaning and wording. Documents are read as they are laid out, in every language: passages follow a document's sections, clauses, definitions, questions or slides; tables, including PowerPoint and Word tables, are kept as tables, and a result can point at the row and the cell; scanned pages are read too. It can read files you already keep elsewhere, or files held in a plycdn bucket. Storage and search are separate: neither requires the other. People can narrow a search step by step - search within the results, up to five times - and choose documents to search only those, then save them as a collection. An assistant on every page of your app answers questions and follow-ups from your library as the answer is written, each claim linked to the moment or page it came from. Each result can bring the passages around it, or its whole section; a file's indexed text can be read back in order; answers can quote their sources word for word, with every quote checked; and a time limit returns the best result in time instead of failing.
- Run it for your own customers. Give each customer of yours a sub-organization with its own usage and limits, create and manage them from your backend code or the dashboard, read your quotas and monthly statements, and see an audit of every change to keys and sub-organizations: Managing sub-organizations from code. The same calls exist in the .NET, Node.js and Python packages (method reference).
- Prove a deletion, limit each tenant. When a file is deleted from search you can fetch a receipt signed by plycdn of what was removed and when its last copies leave our backups, which anyone can check: Proof of deletion. Serving customers of your own, give each a sub-organization and cap its searches, answers and content yourself: Limits for each sub-organization.
2Three steps to a first search
1. Get access and set up your account
Your plycdn contact gives you access to your organization and the values you need to set up callbacks. In the dashboard:
Create your API key under Developers → API keys. It is shown once, so have your secret store ready. Choose its Reach: only the organization selected in the dashboard (the default) or, from your main organization, the whole account. Up to two keys can be active at once for each, so you can replace one without downtime: rotate, deploy the new key, then deactivate the old one, all in the dashboard, which is the only place keys are managed. Keep a standard key in the service your users reach. An admin key, which also manages sub-organizations and reads the account audit from code (
PlycdnAccountApi), and creates, changes or removes storage buckets, video libraries, custom domains, webhook endpoints and the storage account, and rotates signing keys and webhook secrets, belongs only in the automation that sets accounts and buckets up. See "Getting and rotating API keys" in the API reference. Limits are per key and per plan; ask us to raise them.Tell plycdn where your files are, under Content search → Indexing settings. If your files live in your own storage, three things must be in place before your first upload will index. Skip any of them and uploads appear to succeed while nothing is ever indexed.
- Your storage host: the host part of your blob endpoint, for example
yourcompany.blob.core.windows.net. Host name only: nohttps://, port or path. - Your callback URL:
https://YOUR_BACKEND/api/plycdn-content/v1/source-callback. This does not need to be added to your storage host list. - A callback signing secret of at least 32 characters: the same value you configure in your backend as
Plycdn:CallbackSecret.
The ASP.NET Core package does not start without a callback key and secret, and it registers every file with that key, so the callback is part of the integration rather than an optional extra. Your plycdn contact supplies the callback key and confirms your hosts are permitted. The dashboard's Library page shows the state of everything you submit, including
source_host_not_allowedandcallback_not_configuredif any of the above is missing.- Your storage host: the host part of your blob endpoint, for example
The plycdn API key lives only in your backend's secrets. The browser packages use your application's own login and must never receive it.
2. Add the packages to your application
- Backend: ASP.NET Core (.NET 6 and later, up to .NET 10), Node.js (18 and later) or Python (3.9 and later).
Install
Plycdn.ContentSearch.AspNetCore,@plycdn/content-search-nodeorplycdn-content-search, register it, and let it resolve who is signed in and which organization they belong to. It adds the routes your frontend calls, keeps the plycdn key on your server and forwards calls to plycdn. - Frontend: Angular (13 to 22) or React (18 and later). Angular gets independent upload, search and results,
and playback components; React gets hooks (
useSearch,useAnswerand the rest) to build your own interface. Upload, search, results and playback can be placed independently, and themes, labels and result templates are customizable.
Install them from the plycdn feeds with your tool's own command (npm install @plycdn/node, dotnet add package Plycdn.AspNetCore.Storage, pip install plycdn), set up once: Install the SDKs. Or download them from your dashboard: Developers → SDKs, where each language is its own download: pick yours. The Integration guide walks through both sides. Storage and video have their own packages, one guide per platform: ASP.NET Core,
Node.js, Python, Angular and React. The
Node.js and Python guides cover Content Search as well.
3. Upload a file and search it
Upload a recording or a document with the packages' upload component, or register a file you already store. Indexing starts at once; the Library page shows its progress. When it finishes, search for a phrase from inside the file: the result names the moment or page and opens it there. Agree quotas and expected volumes with plycdn before a large first import. Many videos at once? Use Bulk upload on the library's page, or see Moving many videos in.
Prefer to call the HTTP API yourself, with no packages? The API reference covers every endpoint, and the machine-readable description of the Content Search API is published as an OpenAPI file at /docs/openapi.json, for code generators and coding assistants.
Try it in the dashboard first
The dashboard's Content search → Search page is a playground for your own library. Type a question and it shows the
passages found, each with its confidence, its relevance and how it matched (keyword, semantic or both), the
page or moment it came from, and the written answer with its citations. A question the library cannot answer says so
instead of guessing. Content search → Search analytics lists the searches that found nothing, which is the report
of questions your library could not answer.
3What it costs
You pay for what you use, at list prices published on the Plans and prices page, with the
quantities counted as described in "How usage is counted" in the Integration guide. You can
go live on any day of the month: the first and last months are prorated. GetUsageAsync and the usage and
statement methods read what your account has used, by whole month or by day, against your plan's quotas, and read
closed monthly statements including a CSV download. "What you have used" in the API reference is the
field-by-field reference.
4Where to go next
Step-by-step guides
Each takes one task from start to finish: where to click in the dashboard, the code for your server and your pages, how to check it works, and what to do when it doesn't.
| You want to | Guide |
|---|---|
| Make a course that must be watched in order, and read each learner's progress | Serial playback |
| Know how long each learner, member or customer watched, per viewer and per video | Watch time |
| Show each viewer's details over your videos, and your logo, so a leaked recording shows who it came from | Viewer watermark and your logo |
| Go live from an encoder or the browser, on your site or a share link | Live streaming |
| Run ad breaks, live captions, protection, passthrough or a backup encoder on a live stream | Ad breaks, Live captions, Content protection, Passthrough, Backup ingest |
| Send people a link to a stream, in your own colours and on your own address | Your watch page |
| Upload a video and play it in your own application | Playing video |
| Store a video's reference and play it on your own domain | Store and sign on your domain |
| Show files, documents and images in your application: a PDF, a Word or Excel file, a picture, a video, or a search result at the match | Showing any file |
| Keep files private, use your own domain, replace a file, see what each customer used | Delivering files and video |
| Keep files to your own sites and countries, and cap how fast one link downloads | Restrict where files play |
| Upload files and folders in the dashboard, delete several at once, make a link that expires | Manage files in a bucket |
| Serve several buckets, and a video library, from one address of your own, folder by folder | Serve several buckets from one address |
| Redirect a folder or a whole address, or share traffic between buckets with backups, on your own address | Redirect a path, Share traffic between buckets |
| Install the packages with npm, NuGet or pip, and keep them up to date | Install the SDKs |
| Add search, answers and suggestions to your application | Adding search |
| Choose where a bucket's files are stored: Mumbai, Hyderabad or Ashburn (US East) | Choose where files live |
| Keep a second copy of a bucket in another location, in case a region is lost | Keep a disaster-recovery copy |
| See what people search for, and what they searched for and didn't find | See what people search for |
| Be told when a video is ready or a stream starts | Webhooks |
| Sign in without a password, with your fingerprint, face or a security key | Sign in with a passkey |
| Read your bill, see this month's running charges by day, and get them by email or from your server | Your bill and running charges |
| Invite your team, hold your keys, check what a key reaches, serve your own customers | Team, keys and customers |
| Find the organization, bucket, library and other IDs the packages ask for | Find your IDs |
| Bring an existing library across | Moving your library |
| Move your files and your address from Azure, S3 or another web server, without downtime | Moving from another cloud |
| Open course packages and other web pages in the browser on your own domain | Courses and other web pages |
| See whether your tool's requests reach plycdn, and why one was refused | Check that requests reach plycdn |
| Keep your applications' logs, metrics and traces, and search them | Logthread (recommended), or plycdn's built-in Observability; to send plycdn's own activity there, see Logthread |
| Check a query before you run it, keep it within a time limit, see what it read, and see what you sent and what was refused | Check a query, Time limit, What it read, Sent and refused |
| Give a customer, a team or a partner a key that reads only part of your logs, and revoke it at any time | Scoped log access |
| Check that your nightly archive arrived and is intact | See what the archive did |
| Search your storage requests, video events and daily delivery totals in Logthread | Send your plycdn activity to Logthread |
| Let your AI assistant (Claude, ChatGPT, Cursor, Windsurf, VS Code) read these docs and look at your account, by signing in | Connect your AI assistant |
Where things are in the dashboard
| Section | What you do there |
|---|---|
| Content search | Your searchable library, search, collections, relevance tuning, storage connectors, indexing settings, library health and what people search for |
| Storage | Storage buckets (files, uploads, signed links, S3 and Azure access), moves from other clouds and the request log |
| Video streaming | Video libraries (videos, players, serial playback, protection, bulk upload), watch time and your organization's video encoding |
| Live streaming | Live streams (go live, share, restream, ad breaks, captions, recordings) and the watch page for share links |
| Delivery | Delivery analytics and breakdown, caching and CORS, purging |
| Domains & routing | Custom domains and the shared-address routes that send each path to a bucket |
| Monitoring & logs | Logs, metrics and traces: settings, ingest keys, query keys, volumes, slow queries and the archive |
| Developers | API keys, storage credentials, webhooks, integrations, AI assistants and the SDK downloads |
| Billing & usage | Usage, quotas and your plan |
| Account & security | Your profile and organization, team, sub-organizations, branding (player logo, viewer watermark, watch page), sign-in and security, audit log |
Reference
| You want to | Read |
|---|---|
| Connect your application end to end | Integration guide |
| Find the call for a task in .NET, Node.js and Python, side by side | Method reference |
| Create customers' sub-organizations, set their limits, read statements | Managing sub-organizations from code |
| Put files on your own domain, purge caches, read delivery analytics | Storage |
| See plans, what each includes and the list prices | Plans and prices |
| Add storage and video from your backend or your pages | ASP.NET Core, Node.js, Python, Angular, React |
| Judge plycdn against your requirements, capability by capability | Evaluating plycdn |
| Build directly against the HTTP API | API reference, and the OpenAPI file |
| Buckets, uploads, signed links, caching, CORS, purges, usage and delivery analytics, including traffic breakdowns by customer, course or campaign | Storage |
| Video libraries, encoding, the player, the viewer watermark, live streaming and signed webhooks | Video |
| Send logs, metrics and traces, then query, export and tail them | Observability |
| Use S3 and Azure tools with plycdn buckets | S3 and Azure compatibility |
| See what changed in each release, and how to move to it | Release notes and Upgrading |
5Reporting a file
If you believe a file plycdn is storing or delivering should not be, or that a stored file infringes rights or breaks the law, report it to [email protected] with the bucket and object key (or the link) and the reason. The Storage guide describes what happens to a reported file.
For help with an integration, contact your plycdn representative.