Confluence

Confluence is a team workspace and wiki platform designed for creating, organising, and sharing knowledge within organisations. At its core, it's a structured documentation system with pages organised into spaces, supporting rich text editing, embedded media, page hierarchies, templates, and collaborative editing.

The platform includes blog posts for announcements, page comments and inline comments for discussion, tasks for lightweight action tracking, whiteboards for visual collaboration, and databases for structured data. It integrates tightly with Jira (linking issues to documentation), Trello, and other Atlassian products.

Target customers range from small teams (free tier supports up to 10 users) to large enterprises. It's particularly popular with software development teams, IT departments, and organisations already using Jira. The main competitors are Notion, SharePoint, and Google Docs, though Confluence occupies a distinct niche as a structured wiki rather than a general-purpose document editor.

Confluence Cloud uses per-user monthly pricing with volume discounts. The Free tier supports up to 10 users with 2GB storage. Standard starts at roughly $5.16 per user/month (billed annually), with costs decreasing at higher user counts. Premium runs around $9.73 per user/month and adds unlimited storage, 1,000 automation runs per user/month, analytics, and 24/7 support. Enterprise pricing is custom.

For developers and integrators, the API access comes with the platform licence - there's no separate API fee. However, the points-based rate limits scale with the customer's plan tier, so a Free-tier customer gives your integration significantly less API headroom (65,000 points/hour) compared to an Enterprise customer (up to 500,000 points/hour). This matters if you're building a commercial integration that needs to work across different customer tiers.

Data Center licensing (for self-hosted) starts at 500 users with annual licensing, but new Data Center subscriptions won't be available after March 2026, and the product reaches end-of-life in 2029.

Lock-in risk is moderate to high. The primary concern is the Atlassian Document Format (ADF) - all rich content is stored in this proprietary JSON format, and converting it to other formats requires custom tooling. Page hierarchies, macros, inline comments, and embedded Jira references all create dependencies that are hard to replicate elsewhere.

On the positive side, you can export space data via XML/HTML/PDF, and the API allows programmatic extraction of all content. But a full migration to a competing platform (Notion, SharePoint, Google Docs) would require significant effort to preserve formatting and structure.

The bigger strategic concern is Atlassian's push to Cloud-only. With Data Center reaching end-of-life in 2029, organisations lose the option of self-hosting. If you're building integrations for on-prem Confluence customers, those integrations have an expiry date.

Confluence is built by Atlassian, an Australian-founded software company established in 2002 by Mike Cannon-Brookes and Scott Farquhar in Sydney. The company is publicly traded on NASDAQ (TEAM) with a market cap around $22 billion as of early 2026 and annual revenue of approximately $5.8 billion. Atlassian employs around 13,800 people globally.

Atlassian has been on an acquisition spree recently - picking up Loom (video messaging, $975M) in late 2023, Rewatch (AI meeting recorder) in 2024, and three companies in September 2025 alone: Cycle App, The Browser Company, and DX (developer experience, $1B). This aggressive expansion signals they're not going anywhere, but it also means the platform is evolving rapidly.

Confluence itself launched in 2004, giving it over 20 years of market presence. Atlassian discontinued the self-hosted Server product in February 2024 and has announced Data Center end-of-life for March 2029, pushing everyone towards Cloud. This is a deliberate strategic bet on cloud-first that affects how you should think about integrations - Cloud is where the investment is going.

Confluence Cloud offers a RESTful API in two versions: v1 (legacy) and v2 (current). The v2 API covers 26+ resource categories including Pages, Blog Posts, Spaces, Comments, Attachments, Tasks, Labels, Whiteboards, Databases, and Folders. Pagination is cursor-based with a default limit of 50 results and a maximum of 250 per request.

Authentication supports Basic Auth with API tokens for simple scripts, and OAuth 2.0 (3LO) for proper app integrations. Forge and Connect apps use their own auth mechanisms. For OAuth 2.0, access tokens are passed as bearer tokens and the request URL format is https://api.atlassian.com/ex/confluence/{cloudId}/{api}. There's also an Admin Key feature for Premium and Enterprise tiers that provides elevated access.

The big thing to watch is rate limiting. From March 2, 2026, Atlassian enforces a new points-based system. Each API call costs points based on complexity - most GET operations cost 1 point, identity/access queries cost 2 points. Your hourly budget depends on the customer's plan: Free gets 65,000 points/hour, Standard gets 100,000 + (10 x user count), Premium gets 130,000 + (20 x user count), and Enterprise gets 150,000 + (30 x user count), all capped at 500,000. Quotas reset hourly at UTC and unused points don't carry over. Burst limits also apply independently over short windows.

Content is stored in Atlassian Document Format (ADF), a JSON-based rich text format. This is one of the fiddlier parts of the API - you can't just POST plain HTML or Markdown. You need to construct or parse ADF, which has its own learning curve.

Webhooks exist but come with significant caveats. On Confluence Cloud, you can't just register a webhook URL like you would with Stripe or GitHub. You need to build either a Connect add-on or a Forge app and declare webhook subscriptions in your app descriptor. This is a meaningful barrier to entry for simple integrations. Delivery is best-effort with no guaranteed delivery - if your endpoint is down, events are lost. Forge apps have fewer supported webhook events than Connect apps, which is frustrating given that Atlassian is pushing everyone towards Forge (new Marketplace submissions must use Forge from September 2025). Available events cover page/blog CRUD, comment changes, space updates, and attachment operations.

Data export is decent but not seamless. Confluence Cloud supports exporting spaces to XML, PDF, or HTML formats through the admin interface. You can also export database data to CSV. The REST API allows programmatic content extraction - you can fetch pages with their content, attachments, and metadata.

The main lock-in concern is the Atlassian Document Format (ADF). All rich content is stored in ADF, which is Atlassian's proprietary JSON format. While you can extract it, converting ADF to other formats (Markdown, HTML, or another wiki's format) requires custom transformation logic. Page hierarchies, inline comments, and embedded macros add further complexity to any migration.

Import is more limited - you can import from other Confluence instances (Cloud or Server 6.0+), but importing from non-Confluence sources requires building custom solutions via the API. There's no built-in import from Notion, SharePoint, or other competitors.

Documentation quality is solid - around 7/10. The v2 API reference is well-structured with OpenAPI specs available for download and a Postman collection. There are clear examples for common operations. The main pain point is navigating the v1/v2 split - some operations are only documented for v1, and it's not always clear which version you should use for a given task.

SDK support is community-driven rather than official. There's no Atlassian-maintained Confluence SDK. The best option for Node.js is confluence.js (community-maintained, covers both Cloud and Server APIs). Python has confluence-rest-library on PyPI. These work but lack the polish of vendor-maintained SDKs.

For testing, Atlassian provides free Cloud developer instances through the developer portal. You can create a test site with dummy data, which is genuinely useful for integration development. The Atlassian SDK also allows running a local Data Center instance for plugin development, though this is less relevant for Cloud API work.

The developer community is active on the Atlassian Developer Community forums, and Atlassian staff do respond to questions. Stack Overflow coverage is reasonable for common issues. The main frustration developers report is the pace of API changes and repeated deadline shifts for v1 deprecation.