Introducing our new Background Orchestrator

Most research tools hand you an answer. We wanted to give you a process you can actually direct, and then walk away from while it runs. That is what our new background orchestrator does. It formulates and executes complex research workflows, using agents that can run demanding jobs, including biocompute, in the background, so you are not stuck watching a progress bar. Kick off the work, and come back when it is done.

---

The Background Orchestrator Execution Lifecycle

The orchestrator operates through a rigorous, multi-stage lifecycle to ensure scientific accuracy and complete progress traceability:

---

You start with a question, the same way you always would. Instead of jumping straight into execution, the orchestrator clarifies scope with a few follow-up questions and puts together a detailed plan before anything runs: the approach it intends to take, the assumptions it is making, any constraints, how it will validate its own work, what deliverables to expect, and what success actually looks like for the task. You review that plan and approve it before the orchestrator moves forward.

Once approved, the job is handed off and runs in the background. Everything about the process is completely traceable: a reasoning layer tracks the full timeline step by step in real time, along with every file generated along the way, so you can check in whenever you want, or just come back once it is done.

When the job completes, everything lands in one place: structural files, data files, the actual code behind the analysis, and a summary file that ties the whole output together and points you to where each answer lives.

The background orchestrator can also pull directly from the library. That means a complex, multi-step research job is not just reasoning over public literature and generic databases, it can draw on the papers you have already collected, the collections you have organized around a specific question, and the workflows your team has already built. The output ends up directed toward what you actually need.

Complex research questions rarely resolve in one step. They need a plan, real compute time, and they need to be checked against evidence, your own and the field’s. The background orchestrator handles that whole arc: plan, execute, verify, deliver, while staying completely traceable at every step.

Validation is a core part of how we think about this, not an afterthought. A result is only useful if it is verifiable. You should always be able to see how it was generated and exactly where it came from. That is a standard we hold ourselves to across the platform, and it is built directly into how the background orchestrator works.

---

Watch Purna’s Background Orchestrator in Action

Check out our complete overview and hands-on walkthrough video below to see how to formulate, execute, and verify your multi-step pipelines: