Sources
Sources are how content enters a Knoku project. Once indexed, the same project can answer from that content through the widget, bots, MCP, API, support workflows, and dashboard tools.
Open Project -> Sources to add website crawls, GitHub repositories, and file uploads, and to monitor connected sources. Third-party connectors (GitHub App, Notion, Confluence, Jira, OpenAPI, Zendesk Help Center) are configured under Project -> Settings -> Integrations; synced documents still appear in Documents and on Sources once connected.
Choose a source
| Source | Best for | Setup |
|---|---|---|
| GitHub repo sync | Markdown docs in a GitHub repository | Sources -> Add GitHub repository |
| Website crawl | Public HTML docs where the live site is the source of truth | Sources -> Add website crawl |
| GitHub Discussions | Accepted community answers | Settings -> Integrations -> GitHub |
| GitHub Issues & PRs | Workarounds, release context, troubleshooting threads | Settings -> Integrations -> GitHub |
| Notion | Team notes, runbooks, internal knowledge | Settings -> Integrations -> Notion |
| Confluence | Workspace wiki pages and runbooks | Settings -> Integrations -> Confluence |
| Jira | Known issues and reusable ticket context | Settings -> Integrations -> Jira |
| OpenAPI | Endpoint, parameter, and schema reference | Settings -> Integrations -> OpenAPI |
| File uploads | PDF, DOCX, Markdown, and text files | Sources -> Upload |
| Zendesk Help Center | Public support articles by section | Settings -> Integrations -> Zendesk Help Center |
Most projects start with GitHub repo sync or website crawl. Add workspace and support sources after the public docs are working.
GitHub has two source paths
| Path | What it indexes | Where to configure | Guide |
|---|---|---|---|
| GitHub repo sync | Documentation files from a repository | Sources | GitHub repo sync |
| GitHub App integration | Discussions, Issues, and Pull Requests | Settings -> Integrations -> GitHub | Discussions, Issues & PRs |
Use repo sync for formal docs. Use the GitHub App when valuable answers live in community or engineering threads.
Public and internal groups
Every document belongs to one or more source groups. Source groups decide which surfaces can use a document.
| Group | Typical content | Reaches |
|---|---|---|
| Public | Public docs, crawled pages, help center articles, public GitHub community answers | Widget, support form deflector, MCP, Public API keys with public scope |
| Internal | Private repos, Notion, Confluence, Jira, uploaded files, restricted Zendesk articles | Team/support surfaces and Public API keys with internal scope |
Read Source groups before connecting private or internal sources.
Citation URLs
Citations are only useful when they open a real page:
- GitHub repo sync cites the GitHub file URL.
- Website crawl cites the live page URL.
- GitHub Discussions, Issues, and PRs cite GitHub permalinks.
- Workspace sources cite the source page, issue, or article when available.
- Uploaded files do not have public visitor URLs and should stay internal.
If visitors should open a public page from a citation, index content from a public repository or a published website.
Keeping content current
| Source type | Refresh behavior |
|---|---|
| GitHub repo sync | Background sync follows the configured branch; Sync is available from the source row |
| Website crawl | Manual Re-crawl; Business and above can enable automatic refresh |
| Connected workspace/support sources | Re-index or save filters to sync; background workers keep sources current |
| Uploads | Static until replaced or deleted |