Reuse the source.
Your shell, styles, and helpers stay the foundation.
FINE TUNING
Your code.
Your conventions.
At Questra speed.
Give Questra your templates, components, and coding standards. Make every new survey feel like yours.
THE WAY YOU WORK, BUILT IN
Client standard / Decipher01
A familiar foundation.
A new study.
Your survey shellLAYOUT + HELPERS
Your componentsREUSABLE CODE
Your house rulesTEAM PREFERENCES
+ This study’s questions & logic
Unmistakably
your way.
your way.
TEMPLATE ANATOMY
What stays yours.
What Questra adds.
Explore the parts. Add a questionnaire. Watch your code fill in.
Your survey shell. Showing the template.
One named template. Shared across your workspace. Selected for each programming run.
Keep the foundation. Fill in the study.
Questra fills the content slots. Your surrounding source stays the foundation.
BASE TEMPLATEDecipher · XML excerpt
<survey> Your existing source
<style name="respview.client.css">
.question-text { font-weight: 600; }
</style>
<exec when="init">
def is_adult(age):
return age >= 18
</exec>{{slot:screener}}Screening{{slot:main}}Main questionnaire{{slot:closing}}Closing text</survey>Content slots mark where the new study goes.
Shell slots: Decipher and ConfirmIt. Components, instructions, and files: all four supported platforms.
LESS ROOM FOR HALLUCINATION
Less code to invent.
More code you trust.
Supply the implementation. Let Questra fill in the study.
Define the inputs.
The agent supplies values. Your component renders them.
Check the new work.
Code review and link testing check study-specific behavior.
Explore testing & suggestionsReuse reduces invention. New logic still needs review.
GET STARTED
Start with your best work.
Turn an existing survey into your next template.
- 01
Bring an existing survey.
Upload an export or choose a connected survey.
- 02
Extract. Inspect. Refine.
Populate with AI. Review what it derives.
- 03
Choose it for the next study.
Share your named template across the workspace.
One template per standard.
Client A · DecipherClient B · DecipherInternal tracker · ConfirmIt
Fine tuning, explained.
Is fine tuning the same as training an AI model?
In Questra, fine tuning means supplying programming templates, components, and instructions. It does not retrain a model. Your template stays scoped to your workspace and is not shared across workspaces or used to train a Questra model.
How does this reduce hallucinations?
The supplied survey shell and registered components provide existing source code, so the agent does not need to recreate those implementations from a description. Content slots and component inputs define where study-specific values go. Choosing the right components and implementing the questionnaire still require reasoning, which is why code review and link testing remain part of the workflow.
Which survey platforms support fine tuning?
Templates are available for Decipher, ConfirmIt, Qualtrics, and Alchemer. Each template belongs to one platform. Decipher and ConfirmIt also support native base-template assembly using content slots. Qualtrics and Alchemer templates currently carry components, instructions, and supporting files without a slot-compiled native shell.
Can we keep a different template for each client?
Yes. Create a named template for each distinct client or team standard, including multiple templates for the same platform. Templates are shared within your workspace. Choose the appropriate template when you start programming, or explicitly skip it to use Questra defaults.
Do we have to build the template by hand?
You can paste your base source and register components manually, or create a new template and use Populate with AI on an existing survey export or a survey in a connected account. Review what it derives before using it. Populate replaces the base template, components, and instructions on the selected template, so start with a new one when you want to preserve existing work.
YOUR STANDARD. EVERY STUDY.
Make your best work
the starting point.
Bring your programming conventions to Questra.
No credit card required.