Assistant
Realiably recognizes the power of the knowledge available to engineers and leaders today. We have created a unique assistant that brings this mass of information to you
Bring the World in to help you
Reliably allows you to use OpenAI models to ask questions that are relevant to the current experiment. The goal is to provide context when reviewing the execution results.
Reliably never sends your experiment or execution results to OpenAI. Only the questions are passed on.
Leveraging the capabilities of public large language models, such as OpenAI ChatGPT, Reliably contextualizes your executions with well-targeted conversations.
The power of the Reliably Assistant resides in its capacity to enable your creativity by bringing the threads of your execution into a larger context and letting you figure out what to do next from here.
Open up to your creativity
Creating experiments with Reliably is easily achieved through its Builder. Reliably takes it one step further by using the power of LLM to support you generating the right experiment through a simple prompt approach.
The first time you access the assistant builder, it will ask you for an OpenAI key. That key will be encrypted and stored into Reliably’s database.
Reliably uses the GPT-4.1 model. Please review its costs.
Once configured, you can start using the assistant builder by passing a prompt:
Select now the target platform:
Once you press enter, Reliably Assistant will start generating an experiment for you:
Reliably will then ask you for a few parameters:
Finally, give the experiment a meaningful title and it’s done!
The experiment is now created and is part of your catalog. You may decide to run it on the spot, edit it or view it. Let’s see what happens when we run it:
Select the appropriate deployment runner, the right environment and off you go!
Reliably will then schedule the epxeriment:
Once completed, Reliably will show you a link back to the execution’s results:
