Knowledge - SiteGPT Docs

Knowledge commands manage the content SiteGPT uses to answer questions. Run this first when exploring:

sitegpt knowledge --help

Knowledge resources

Resource	Purpose
`knowledge documents`	List, inspect, edit, resync, delete, and measure ingested documents. Alias: `knowledge docs`.
`knowledge links`	Add one or more exact URLs.
`knowledge website`	Crawl a website from a starting URL.
`knowledge sitemap`	Add URLs from a sitemap.
`knowledge youtube`	Add YouTube video URLs.
`knowledge text`	Replace the chatbot text knowledge document.
`knowledge files`	Upload local files.
`knowledge sync-jobs`	Manage recurring content sync and sitemap scan jobs.
`knowledge sources`	Manage external data-source connections. Alias: `knowledge connections`.
`knowledge custom-responses`	Manage dashboard Custom Responses.

Add knowledge

Add exact links

Use this when you know the exact pages to train:

sitegpt knowledge links add --chatbot <chatbot-id> https://example.com/a https://example.com/b
sitegpt knowledge links add --chatbot <chatbot-id> https://example.com/a --only-main-content true
sitegpt knowledge links add --chatbot <chatbot-id> https://example.com/a --sync WEEKLY

Crawl a website

Use this when you have a starting URL and want SiteGPT to discover pages:

sitegpt knowledge website add --chatbot <chatbot-id> https://docs.example.com
sitegpt knowledge website add --chatbot <chatbot-id> https://docs.example.com --depth 3 --max-links 200
sitegpt knowledge website add --chatbot <chatbot-id> https://docs.example.com --include-path /docs --exclude-path /blog
sitegpt knowledge website add --chatbot <chatbot-id> https://docs.example.com --allowed-domain docs.example.com

Website options:

Option	Description
`--depth <1-5>`	Maximum crawl depth. Default: `3`.
`--max-links <1-1000>`	Maximum URLs to queue. Default: `50`, capped by remaining quota.
`--include-path <path>`	Include path pattern. Repeatable.
`--exclude-path <path>`	Exclude path pattern. Repeatable.
`--allowed-domain <domain>`	Allowed domain. Repeatable.

Add a sitemap

Prefer sitemap ingestion when the website exposes a sitemap:

sitegpt knowledge sitemap add --chatbot <chatbot-id> https://example.com/sitemap.xml
sitegpt knowledge sitemap add --chatbot <chatbot-id> https://example.com/sitemap.xml --max-links 500
sitegpt knowledge sitemap add --chatbot <chatbot-id> https://example.com/sitemap.xml --scan WEEKLY --sync WEEKLY

Sitemap options:

Option	Description
`--max-links <1-1000>`	Maximum URLs to queue. Default: `50`, capped by remaining quota.
`--include-path <path>`	Include path pattern. Repeatable.
`--exclude-path <path>`	Exclude path pattern. Repeatable.
`--sync <frequency>`	Refresh known page content. Values: `NEVER`, `DAILY`, `WEEKLY`, `MONTHLY`. Default: `NEVER`.
`--scan <frequency>`	Scan sitemap for added/deleted URLs. Values: `NEVER`, `DAILY`, `WEEKLY`, `MONTHLY`. Default: `NEVER`.

Sync and scan frequencies are plan-gated. Unsupported requests may be downgraded by the API with warnings.

Add YouTube videos

sitegpt knowledge youtube add --chatbot <chatbot-id> https://www.youtube.com/watch?v=...
sitegpt knowledge youtube add --chatbot <chatbot-id> https://www.youtube.com/watch?v=... --sync MONTHLY

Upload local files

sitegpt knowledge files add --chatbot <chatbot-id> ./guide.pdf
sitegpt knowledge files add --chatbot <chatbot-id> ./pricing.pdf ./faq.docx

The CLI reads local files and uploads them. Pass file paths, not remote URLs.

Add or update text knowledge

text update creates or replaces the chatbot’s text knowledge document:

sitegpt knowledge text update --chatbot <chatbot-id> "Refund policy: ..."
sitegpt knowledge text update --chatbot <chatbot-id> --name FAQ --file ./faq.md

text add is accepted as an alias, but text update is clearer because the text block is replaced.

Scrape options

These options work with links, website, sitemap, and document config updates:

Option	Description
`--only-main-content <boolean>`	Extract main content only. Accepts `true`, `false`, `1`, `0`, `yes`, `no`, `on`, `off`. Default: `true` for new URL ingestion.
`--include-selector <selector>`	CSS selector to include. Repeatable.
`--exclude-selector <selector>`	CSS selector to exclude. Repeatable.
`--header "Name: value"`	Send request header. Repeatable. Allowed names: `User-Agent`, `Accept`, `Accept-Language`.
`--sync <frequency>`	Auto-sync page content. Values: `NEVER`, `DAILY`, `WEEKLY`, `MONTHLY`.

Example:

sitegpt knowledge links add \
  --chatbot <chatbot-id> \
  https://example.com/docs \
  --only-main-content true \
  --include-selector main \
  --exclude-selector nav \
  --header "User-Agent: SiteGPT"

Documents

Documents are individual trained, pending, failed, or queued knowledge items.

List documents

sitegpt knowledge documents list --chatbot <chatbot-id>
sitegpt knowledge documents list --chatbot <chatbot-id> --json
sitegpt knowledge documents list --chatbot <chatbot-id> --status SUCCESS
sitegpt knowledge documents list --chatbot <chatbot-id> --source LOCAL_FILE
sitegpt knowledge documents list --chatbot <chatbot-id> --state failed
sitegpt knowledge documents list --chatbot <chatbot-id> --query "refund"

knowledge docs is accepted as an alias for knowledge documents. Filters:

Option	Description
`--query <text>`	Search by document name/content.
`--source <source>`	Filter by source. Repeatable.
`--status <status>`	Filter by ingestion status. Repeatable.
`--type <type>`	Filter by document type. Repeatable.
`--limit <1-100>`	Limit results. Default: `50`.
`--cursor <cursor>`	Continue from a previous response.

Allowed source values:

TEXT, URLS_LIST, YOUTUBE, SITEMAP, WEBSITE, LOCAL_FILE,
NOTION, GOOGLE_DRIVE, DROPBOX, ONEDRIVE, BOX, SHAREPOINT,
CONFLUENCE, GITHUB

Allowed status values:

BACKLOG, QUEUED, QUEUED_FOR_RESYNC, QUEUED_FOR_UPDATE,
QUEUED_FOR_DELETION, PROCESSING, SUCCESS, FAILED, CANCELLED

Allowed type values:

TEXT, URL, FILE, NOTION_DOCUMENT, GOOGLE_DRIVE_DOCUMENT,
DROPBOX_DOCUMENT, ONEDRIVE_DOCUMENT, BOX_DOCUMENT,
SHAREPOINT_DOCUMENT, CONFLUENCE_DOCUMENT, GITHUB_DOCUMENT

Get document details

sitegpt knowledge documents get --chatbot <chatbot-id> <document-id>
sitegpt knowledge documents get --chatbot <chatbot-id> <document-id> --content
sitegpt knowledge documents get --chatbot <chatbot-id> <document-id> --include-content --max-content-chars 50000 --json

--content includes raw and parsed text when available. Use --max-content-chars <1-500000> to limit included text.

Edit document content

sitegpt knowledge documents edit --chatbot <chatbot-id> <document-id> --content "Updated content"
sitegpt knowledge documents edit --chatbot <chatbot-id> <document-id> --file ./updated-content.md

documents update is accepted as an alias for documents edit.

Resync documents

sitegpt knowledge documents resync --chatbot <chatbot-id> <document-id>
sitegpt knowledge documents resync --chatbot <chatbot-id> --document <document-id>
sitegpt knowledge documents resync --chatbot <chatbot-id> --state failed
sitegpt knowledge documents resync --chatbot <chatbot-id> --source WEBSITE --status FAILED

Delete documents

sitegpt knowledge documents delete --chatbot <chatbot-id> <document-id> --yes
sitegpt knowledge documents delete --chatbot <chatbot-id> --document <document-id> --yes
sitegpt knowledge documents delete --chatbot <chatbot-id> --state failed --yes
sitegpt knowledge documents delete --chatbot <chatbot-id> --all --yes

documents remove is an alias for documents delete. Bulk selector options:

Option	Description
`--document <id>`	Select a document ID. Repeatable.
`--document-id <id>`	Alias for `--document`. Repeatable.
`--state <state>`	Values: `all`, `trained`, `pending`, `failed`.
`--all`	Select every matching document. Use carefully.
Filters	`--query`, `--source`, `--status`, and `--type` can narrow bulk operations.

Update document config

sitegpt knowledge documents update-config --chatbot <chatbot-id> --document <document-id> --sync WEEKLY
sitegpt knowledge documents update-config --chatbot <chatbot-id> --document <document-id> --only-main-content false
sitegpt knowledge documents update-config --chatbot <chatbot-id> --document <document-id> --include-selector main
sitegpt knowledge documents update-config --chatbot <chatbot-id> --document <document-id> --exclude-selector nav
sitegpt knowledge documents update-config --chatbot <chatbot-id> --document <document-id> --header "User-Agent: SiteGPT"

Document stats

sitegpt knowledge documents stats --chatbot <chatbot-id>
sitegpt knowledge documents stats --chatbot <chatbot-id> --source LOCAL_FILE

Sync jobs

Sync jobs are recurring maintenance jobs. Auto-sync refreshes known content. Sitemap auto-scan discovers URL additions/deletions.

sitegpt knowledge sync-jobs list --chatbot <chatbot-id>
sitegpt knowledge sync-jobs get --chatbot <chatbot-id> <job-id>
sitegpt knowledge sync-jobs update --chatbot <chatbot-id> <job-id> --sync WEEKLY
sitegpt knowledge sync-jobs update --chatbot <chatbot-id> <job-id> --scan MONTHLY
sitegpt knowledge sync-jobs delete --chatbot <chatbot-id> <job-id> --yes

Aliases:

knowledge jobs → knowledge sync-jobs
knowledge sync-jobs disable → knowledge sync-jobs delete

List filters:

Option	Values
`--source <source>`	`URLS_LIST`, `SITEMAP`, `WEBSITE`, `YOUTUBE`, `GITHUB`; repeatable
`--sync <frequency>`	`DAILY`, `WEEKLY`, `MONTHLY`; repeatable
`--scan <frequency>`	`NEVER`, `DAILY`, `WEEKLY`, `MONTHLY`; repeatable
`--limit <1-100>`	Default: `50`
`--cursor <cursor>`	Continue pagination

Deleting a sync job disables recurring sync/scan. It does not delete existing knowledge documents.

Custom responses

Custom Responses are manual question-and-answer overrides in the Knowledge section.

sitegpt knowledge custom-responses list --chatbot <chatbot-id>
sitegpt knowledge custom-responses list --chatbot <chatbot-id> --state OPEN
sitegpt knowledge custom-responses list --chatbot <chatbot-id> --source MANUALLY_ADDED
sitegpt knowledge custom-responses get --chatbot <chatbot-id> <custom-response-id>
sitegpt knowledge custom-responses add --chatbot <chatbot-id> --question "What is your refund policy?" --answer "..."
sitegpt knowledge custom-responses update --chatbot <chatbot-id> <custom-response-id> --answer "Updated answer."
sitegpt knowledge custom-responses delete --chatbot <chatbot-id> <custom-response-id> --yes

Filters:

Option	Values
`--query <text>`	Search questions.
`--state <state>`	`ALL`, `OPEN`, `UPDATED`. Default: `ALL`.
`--source <source>`	`DOWNVOTED_BY_USER`, `MANUALLY_ADDED`, `MARKED_BY_ADMIN`. Repeatable.

Fields:

Option	Description
`--question <question>`	Custom response question.
`--answer <answer>`	Custom response answer.
`--original-answer <answer>`	Optional original answer when adding a custom response.

​Knowledge resources

​Add knowledge

​Add exact links

​Crawl a website

​Add a sitemap

​Add YouTube videos

​Upload local files

​Add or update text knowledge

​Scrape options

​Documents

​List documents

​Get document details

​Edit document content

​Resync documents

​Delete documents

​Update document config

​Document stats

​Sync jobs

​Custom responses

​Full reference

Knowledge resources

Add knowledge

Add exact links

Crawl a website

Add a sitemap

Add YouTube videos

Upload local files

Add or update text knowledge

Scrape options

Documents

List documents

Get document details

Edit document content

Resync documents

Delete documents

Update document config

Document stats

Sync jobs

Custom responses

Full reference