The Forge LLMs API is now in Preview, allowing developers to call Claude models (Haiku, Sonnet, Opus) directly within Atlassian's Forge platform without data leaving the Atlassian environment. Hosted on AWS Bedrock, the API requires minimal setup — just a few lines of code using the @forge/llm SDK. Apps using it can qualify for the 'Runs on Atlassian' badge, making it suitable for enterprise and government customers with strict data governance requirements. Usage is billed in credits based on token consumption, starting June 1, 2026. Example use cases include Confluence AI summarization, agentic LLM web triggers, and long-running LLM processes with real-time UI updates.
Table of contents
Try it outModel availabilityInspiration and examplesPricing and observabilityReady to start building?245 Impressions