AI Playground

Testing prompts, comparing models and trying every generation type in the Aimogen Pro Playground before committing to a rule.

The Playground runs any generation type in isolation, with no rule machinery around it. It is the right place to develop prompts.

Advanced Tools › AI Playground

Why use it#

Tuning a prompt by running bulk rules is slow, expensive and confounded by everything else the rule does. The Playground removes all of that: one prompt, one model, one result, immediately.

Move the prompt into the rule once it works.

The tabs#

Model Comparison#

The most useful tab. One prompt, several models, results side by side.

It answers the question you actually have when output is disappointing: is this the prompt or the model? If every model produces the same problem, fix the prompt. If only one does, change the model.

It is also the honest way to justify model cost. Run your real prompt on the expensive model and the cheap one, and look at whether the difference is worth the multiple.

Prompt Library#

Saves prompts for reuse. Worth using: prompts that work are assets, and they get lost otherwise.

Cost#

Playground requests are real API requests, billed normally and recorded in System & Logs › Usage Logs. Comparing four models on one prompt is four requests.

That is still far cheaper than tuning by running bulk rules.

Access#

The Playground requires the access_aiomatic_menu capability, and it can be hidden through Workspace or menu restrictions.

Related#

• Prompt best practices
• Debugging prompts
• Selecting models