Kirava
Describe the task. Get the automation.
6 followers
Describe the task. Get the automation.
6 followers
Zapier, Make and n8n give you a canvas; you build the flow and maintain it. Kirava generates it. Input is a plain-language description. Output is a runnable artifact Python/Node code, an importable n8n workflow, or a Make blueprint, each with a setup guide. Execution is hosted: the delivery runs in the portal behind a trigger, and each run returns a result card. Revisions are described rather than rebuilt. Metering is on actual token consumption per generation, not per task fired.
One thing that would make this way more useful for me: let me paste in an existing Zapier or n8n workflow and have Kirava reverse-engineer the plain-language description from it. That way I could finally understand what my own three-year-old zaps actually do, and tweak them by editing the description instead of hunting through nodes.
@alfiesolomons Thank you for such a fantastic suggestion! That hits the nail on the head regarding what makes older automations a nightmare where they turn into digital haunted houses that no one dares to touch.
Being able to import JSON from Zapier or n8n to get a 'reverse-engineered' description (and eventually tweak the logic directly through text) is a brilliant idea that would solve that exact problem.We are taking this straight into our product planning. Thank you so much for sharing! Best Victor