Guides
Governed literature-review walkthrough
A literature-review request should state what will be compared and how supporting evidence will be assessed. This SAVRN Cloud guide uses a fictional research task to connect that scope with source checks, an evidence table and a reviewer decision, without presenting fixture text as research.
SAVRN Cloud · Coming soon · Public preview · Browser-local simulation · No live compute or payments
Define the request
Create a project such as “Methods workshop” in Projects. In Work Orders, choose New work order and describe a fictional literature-comparison task. Select an Approved source collection, enter the Expected deliverable and Maximum sample credits, then choose Submit for approval.
A useful sample request is: “Compare three fictional scheduling studies, identify their assumptions, and produce a one-page evidence table.” Specify that the reviewer expects a claim, supporting source and limitation in each row. This tests a complete deliverable rather than a vague prompt.
Follow the local decisions
The new record starts Awaiting approval. Open Approvals, inspect its scope and choose Approve scope and budget to make it ready. Return to Work Orders and choose Run workflow. Inspect the completed sample artifact and Usage, then choose Accept deliverable or Return for revision. Returning requires a note; a revised scope needs a new Work Order.
The generated artifact is a fixture. It is not a real literature search and its content must not be cited as research evidence.
Before connected services launch
Use approved, accessible sources and preserve exact provenance. Verify that each cited passage supports its associated claim. Record exclusions, uncertainty and reviewer corrections. Enforce project access and spending limits before execution. Retain the accepted artifact and accounting receipt under the applicable policy. A citation-looking string or a successful model response is not proof that the literature was reviewed.
Build toward the work that matters.
Tell SAVRN what your institution needs to run, who reviews the results and where its data must stay. That workload defines the next service to qualify.
Discuss an AI project