Add content to Knowledge

Keep the documents and website pages your team needs in Knowledge. Start with a folder, add content, then check that Search can find it.

Upload files

1. Open Knowledge → Browse in the project, workspace or organization that will own the files.
2. Select a folder. To keep a set of documents together, open its actions menu and choose Create subfolder.
3. Choose Searchable for content to index. Choose Raw storage for files to store without adding them to search.
4. Open the folder’s actions menu and choose Upload files. Drop files into the upload dialog or choose them from your computer.
5. Open a file and check its indexing state. Follow Check indexing and search before relying on the content.

To write a short document here, choose Create text file from the folder’s actions menu. Open the file, choose Edit, enter the text and save it.

Empty files wait for content before indexing. If someone changes a file while you edit it, the conflict alert keeps your draft. Choose Reload to inspect the latest version, then Restore my draft to bring your text back for review before saving. Your recovery draft stays in this page. Copy it before leaving.

Add a website

1. Open Knowledge → Sources and add a source. Or choose Add new source here from a folder’s actions menu to select its destination first.
2. Choose Web crawl, name the source and enter the website URL.
3. Choose a reading level. The form offers Basic, Standard and Premium. Reading level is fixed when the source is created.
4. Open the advanced fields to bound the crawl: maximum depth, maximum pages, and URL patterns to include or exclude.
5. Choose how to name the files and a schedule. Use Manual while checking a new source. Save it.
6. Open the source to inspect its runs, URLs and files. Use Run now for a manual refresh.
7. Open a resulting file and search for a fact from it.

Set a small maximum page count for the first run. A source can read and process many pages on later runs. An interval schedule has a minimum and maximum interval, shown in the form and in the limits below.

A failed fetch, including an unreachable hostname or a provider fetch error, makes the run incomplete. Intentional skips caused by crawl filters or robots rules remain separate from fetch failures. Open the source’s URL details to read the fetch error. Existing content stays in place when the crawl cannot check every page. After a successful refresh, check the resulting files and a search result.

Copy or sync files from another tool

Use an S3-compatible client to copy files or sync a local folder. Connect tools over S3 covers access keys, the AWS CLI and rclone. Choose a key for the workspace or project that should receive the files.

After a copy or sync, check the files in Browse and test Search. Stored files and searchable content are different states.

Limits

These are the defaults in the limits catalog. Organization overrides can change the values marked Yes. The source form can also set a lower maximum page count.

Limit	Default	Support can change it
Largest file that can be uploaded: knowledge, the files API, the S3 endpoint, and chat attachments (chat widgets may set a lower value)	25 MB	Yes
Most pages one web source may read in a single crawl run (a source’s own max_pages above this runs at this value)	10,000	Yes
Shortest time allowed between scheduled runs of a knowledge source (interval and cron schedules)	3,600 s	Yes
Longest interval schedule a knowledge source can use	2,592,000 s	No
Requests per minute an organization may make to each public FlowFS API endpoint class (a burst of this size, refilled evenly)	60	Yes
Bytes an organization may upload per day through the public FlowFS API (/v1/files)	5,120 MB	Yes
Longest knowledge source, folder or file name	200	No
Longest knowledge source or folder description	1,000	No
Longest knowledge search query	1,000	No
Longest knowledge access key name	120	No

The daily public API upload allowance applies to the files API. Dashboard uploads don’t use that allowance.