Julep

Julep is a platform for building AI agents that remember past interactions and carry out complex, long running tasks. It targets workflows that go beyond a linear chain of prompts and API calls, such as processes with many steps, decisions based on model output, parallel branches, heavy tool use and long run times. Agents keep context across interactions, and stateful sessions track conversation history and can be paused and resumed. Tasks are defined as multi step workflows with conditional logic, loops and parallel processing, and the platform manages retries, message resending and error handling for long running work. Agents can use built in tools and integrate with external APIs. A document store provides retrieval augmented generation over a team's own data, and SDKs are available for Python and Node.js. The documentation also announces an Open Responses API in alpha, and publishes a machine readable index of its docs. The documentation is organized for AI developers, who start with a quick start guide and core concepts, and for enterprise teams. The listing points to the product documentation rather than a marketing site. The documentation does not publish prices.