Confluence
Turn Confluence spaces into retrievable agent knowledge
- Publisher
- By connic
- Syncs
- Pages, PDF attachments, Labels
- Set up in
- Retrieval → Sources
- Documentation
- Retrieval Docs
How the Confluence source works
Confluence is where engineering wikis, architecture decisions, and internal handbooks live. The Confluence source syncs whole spaces into connic retrieval, so support and operations agents can cite the wiki page instead of improvising an answer.
Connect with a site URL and an API token, pick the spaces to sync, and optionally narrow the scope by label. connic converts each page to text, chunks and embeds it, and keeps it fresh on the configured schedule. PDF attachments can be pulled through the same extraction pipeline used for direct file uploads.
What gets synced
- Pages
- Published page content per space, converted to text with formatting noise stripped.
- PDF attachments
- Optionally ingested through the same document extraction pipeline as direct uploads.
- Labels
- Use labels as a filter so only curated pages reach the retrieval.
Set up the Confluence source
Add the source under Retrieval → Sources in the environment your agents run in. Each environment keeps its own sources.
Create an Atlassian API token at id.atlassian.com/manage-profile/security/api-tokens and copy it (it is shown only once).
Atlassian tokens now expire (up to one year), so plan to rotate it before it lapses.
Enter the Confluence Cloud site URL including the /wiki path, e.g. https://example.atlassian.net/wiki.
Enter the Atlassian account email the token belongs to, then paste the token below.
The sync reads only what that account can see, so give it access to the spaces to ingest.
Configuration, in depth
Every available control for connecting Confluence to a project.
Connection
- Site URL*
- The Confluence Cloud site.
- Account email*
- Email of the Atlassian account the API token belongs to.
- API token*
- Atlassian API token. The sync only reads content this account can read.
Content scope
- Space keys
- Comma-separated space keys, e.g. ENG, OPS. Leave empty to sync every space the account can read.
- Label filter
- Only sync pages carrying one of these labels. Leave empty to sync all pages in scope.
- Include archived pages
- Also sync pages that have been archived in Confluence.
- Include PDF attachments
- Ingest PDF attachments through the document extraction pipeline.
Sync behavior
Where synced content lands and how often it refreshes.
- Target namespace
- Retrieval namespace the synced entries are filed under. Use dots to nest, e.g. sources.notion.
- Sync interval
- How often the source is checked for changes. Each run only reprocesses content that changed.
- When content is deleted at the source
- Pages deleted or moved out of scope in Confluence can be removed from the retrieval on the next sync, or kept as they were last synced.
Search it from an agent
Agents search synced content with the retrieval_query tool. A small tool that pins the namespace keeps an agent to this source.
from connic.tools import retrieval_query
async def search_confluence(question: str) -> dict:
return await retrieval_query(
question, namespace="sources.confluence"
)version: "1.0"
name: support-agent
type: llm
model: connic/gemini-3.7-flash
system_prompt: |
Answer only from what search_confluence returns.
tools:
- knowledge.search_confluenceConfluence retrieval source FAQ
More retrieval sources
Notion
By connicConnect a Notion workspace and keep pages and databases synced in retrieval. Agents retrieve the current version, not a stale export.
Auto-syncSuperhuman Docs (Coda)
By connicSync Superhuman Docs pages and table rows into retrieval, so agents answer from the doc the team actually edits.
Auto-syncWebsite Crawler
By connicPoint the crawler at documentation, a help center, or a blog. Pages are converted to clean markdown and re-crawled on the configured schedule.
Auto-sync