PromptGruup - Quickly iterate LLM workflows to streamline your prompting
by•
Utilize collaborative, node-based interfaces to create effective and precise prompts effortlessly. Collaborate on docs with your team and reduce your time to deployment. Export prompt chains in JSON and YAML to integrate later.
Replies
I built PromptGruup while making another project of mine, ClickCV, and am excited to open it up for the beta launch!
Whether you're integrating LLMs into personal projects, freelance work, or AI-driven applications, structuring effective prompt chains can become tedious. PromptGruup provides a node-based interface that lets you visually design, test, and refine workflows without losing track of what works. Run messages across multiple LLMs with varying parameters at once, pass or parse responses between LLM API calls, and save templates to quickly return to configurations that have worked for you.
PromptGruup is a bloat free solution to prompt chaining. If you're using OpenAI or Anthropic APIs while building your own projects, please check try it out and let me know what you think!
PromptGruup structures its node graphs using a parent-child hierarchy. When composing messages for a new or existing node, you can embed the IDs of any parent nodes whose responses you want to pass in. These embedded IDs will be substituted with the corresponding parent node responses during processing.
If you need to modify or extract specific data from a response before substitution, you can configure a parsing function to the node. Parsing functions are pre-defined with a responses object as the parameter, enabling easy access to parent node responses in the responses.parentNodeId fields. These functions are compiled and ran in their own environments before continuing to fetch LLM responses. Right now, both JavaScript and Python parsing functions are supported.