Your Files Feed

The your files feed makes the files you already keep in a media store searchable. Point it at a store — or one folder in it — and every PDF and picture in there becomes a record you and your agents can search. Nothing to export, nothing to upload again, no copy of your data anywhere new: the files stay exactly where they are.

What makes it different from the other feeds is that it doesn’t reach outside your account at all. There is no URL to give us, no token to paste, and nothing to authorise — it reads the store you already connected, using the credentials you already gave it. This page covers what’s specific to this feed. For how feeds work in general — syncing, binding a list, deleting — see the Feeds reference.

Make your files searchable

1. Go to Feeds (under Sources in the sidebar), click Connect feed and choose Your files.
2. Give the feed a name and pick the media store holding your files.
3. Optionally give a folder — everything in it, including its sub-folders, is included. Leave it blank for the whole store.
4. Choose which files to include: documents, images, or both.
5. Click Connect. We look in the folder there and then, so a wrong folder or one with nothing searchable in it is caught immediately rather than failing quietly later — and the first sync starts right away.

One store does everything

This feed uses a single media store: it reads your files from it, and it writes its own synced file (feeds/<name>.ndjson) back into it. There is no second store to choose.

That matters for one reason, and it’s the thing to get right: the list you import the feed into must use that same media store. Each record points at its file by its path within the store — handbooks/policy.pdf, not a full web address — so we need to know which store to look in, and we take that from the list. Point the list at a different store and the records import but the files behind them can never be found.

In practice this looks after itself: import the feed’s file into a list the normal way and the list uses the store that file came from. It only goes wrong if you deliberately point the list somewhere else.

Using paths rather than web addresses is also what keeps the feed durable. A private bucket can only be read through a temporary signed link, and a link like that would go stale inside your records within hours. A path never expires.

What we can read

Two kinds of file are searchable today:

• Documents — .pdf. We read the words inside the document, so a search matches what a page actually says, not just its file name.
• Images — .png, .jpg, .jpeg. These are searched by what they show, using your list’s model.

Everything else in the folder — Word files, spreadsheets, text files, videos, archives — is skipped for now, quietly and harmlessly. We’d rather leave a file out than add a record that looks searchable but isn’t: a record whose contents we can’t read would match nothing and only get in your way.

A note on images. Whether an image is searchable depends on your list’s model — a model that only understands text can’t look at a picture. See Image Search.

What a record looks like

One file becomes one record:

• Name — the file’s own name, e.g. policy.pdf.
• file — where the file lives in the store. This is what lets a result open the real thing.
• path — the full path, so you can filter by folder.
• type — document or image, so you can filter to one or the other.

We don’t copy what’s inside your files into the record. The record points at the file, and the file itself is read when your list is built — so the searchable content is always the file as it stands, and your data isn’t duplicated into a second place.

Keeping up with changes

On each sync we look at what’s in the folder and compare it with last time. Add a file and it becomes searchable; change one and its record updates; delete one and its record goes away. Nothing to tell us, nothing to re-upload.

New files are included by default — that’s the point of pointing at a folder rather than listing files by hand. Renaming a file counts as removing one and adding another, so it comes back as a new record.

How quickly a change shows up depends on the feed’s sync interval (30 minutes by default, 5 at the fastest). Free accounts pause automatic syncing after a spell of inactivity and resume on sign-in; syncing by hand is never paused. The Feeds page covers both.

How much it can hold

One feed carries up to 5,000 files. If your folder has more than that we tell you when you connect, rather than quietly indexing some of them and leaving you to wonder which — point the feed at a narrower folder, or leave out what you don’t need to search. Several feeds can read the same store, so a big store can be covered folder by folder.

How many records your account can hold in total is set by your plan; see Pricing.

A worked example

Say your team keeps its handbooks and policies as PDFs in a Cloudflare R2 bucket, under handbooks/. You connect that bucket as a media store, add a Your files feed pointed at handbooks/, and import the feed’s file into a new handbooks list with keep-in-sync switched on.

Now “what’s our parental leave policy?” finds the right PDF — because we read what’s written inside it, not just its file name. When someone drops a new policy into the folder, it’s searchable at the next sync, and nobody has to remember to tell the search about it.