Hey Product Hunt!
I built PromptForge because I was tired of the endless trial-and-error cycle of prompt engineering. As someone who works with AI daily, I found myself constantly tweaking prompts, losing track of what worked, and having no systematic way to evaluate improvements.
The problem: Most prompt tools are just fancy text editors. You write, you test, you hope for the best. But there's no real engineering discipline.
What makes PromptForge different:
Systematic evaluation - Generate comprehensive test suites automatically
Built-in analytics - Track what actually works across different scenarios
Variable testing - Test edge cases, robustness, and consistency
Prompt library - Never lose a working prompt again
Dual analysis - AI-powered feedback on your prompts before you even test them
Why I built it: I wanted to bring the same rigor to prompt engineering that we have in software engineering - version control, testing, systematic improvement.
Currently supporting Claude, GPT-4, and Azure OpenAI with more providers coming soon. Docker deployment makes it dead simple to get started.
What would you want to see in a prompt engineering tool? Always looking for feedback from fellow AI builders!
This is cool — would love to be able to run it locally and use @LM Studio or @Ollama for inference. Would that be possible? (I also have Docker installed but it'd be nice to avoid unnecessary $$ cloud bills)
@chrismessina Thanks, Love this feedback! Local inference is definitely on the roadmap!
The Docker setup already runs locally - you just need to plug in your API keys. But adding LM Studio/Ollama support would be AMAZING for cost-free local inference!
Would you prefer Ollama integration first? And what models are you running locally? Always looking to prioritize based on real usage!
@khiemlee Thank you! 🚀 It's been an incredible journey building this. I'll surely checkout what you are releasing . Btw what kind of prompts do you work with most? Code, creative writing, data analysis? Always looking to improve the tool based on real use cases! ✨
Report
@insaanimanav I played with this a little bit the UI is a bit confusing to me. After you generate the evals, I was expecting to be able to save the tests and re-run them but I couldn't figure out how to do that. Is that supported?
I can see the future where every second person tomorrow will be the prompt engineer. So product like promptforge is the need for today .
I would be love to try and be a user for promptforge ❤️
@sumitgoel Love this perspective! You're absolutely right - prompt engineering is becoming a core skill across industries!
Just like how everyone learned Excel or basic coding, systematic prompt engineering will be essential. That's exactly why I wanted to make it accessible with AI assistance and structured testing - democratizing what's currently an art form!
Would love to have you as a user! Try the demo and let me know what you think - especially curious about your experience with the AI-guided prompt generation vs the testing features!
@peter_zhou_overflow That's exactly what I built PromptForge for! What other tools were you testing against? And what made the biggest difference - the AI-assisted generation, the speed, or something else?
Report
Congratulations on the launch, this looks amazing. Curious if you built this yourself with AI?
Yes, I built this myself with AI assistance for parts of the development - felt very meta using AI to build AI prompt engineering tools!
Used Cursor for some of the frontend components, but the core architecture, evaluation algorithms, and product vision were all human-designed. AI was more like a really smart pair programming partner!
The irony wasn't lost on me that I was prompting AI to help build a tool for better prompting!
What's your experience with AI-assisted development? Always curious how other builders are using it in their workflow!
This product really highlights in the built-in analytics and ability to automatically generate test suites. It’s inspiring to bring real software engineering principles to AI prompts.
@pulin_yuuu Thank you! You totally nailed the core philosophy - treating prompts with the same rigor we use for code!
The automated test suite generation has been the biggest game-changer. Just like we wouldn't ship code without unit tests, why ship prompts without systematic validation? The analytics help you understand what's actually working vs just guessing!
Are you coming from a software engineering background? Would love to hear your thoughts on other SE principles we could apply to prompt development - version control, CI/CD for prompts, etc.!
Really appreciate the thoughtful feedback!
Report
People say prompt engineering is not needed since AI is smarter than ever but real users know it’s still a challenge
Congrats on the launch!
Replies
Olvy
@insaanimanav nice work. Good luck.
Smart name tag.
Olvy
@django_vinci Haha thanks
@insaanimanav awesome idea, thanks for sharing. when will this be ready for actual use?
Olvy
@the_league Its out right now you can clone the repository and use the setup instructions to start using it
Raycast
This is cool — would love to be able to run it locally and use @LM Studio or @Ollama for inference. Would that be possible? (I also have Docker installed but it'd be nice to avoid unnecessary $$ cloud bills)
Olvy
@chrismessina Thanks, Love this feedback! Local inference is definitely on the roadmap!
The Docker setup already runs locally - you just need to plug in your API keys. But adding LM Studio/Ollama support would be AMAZING for cost-free local inference!
Would you prefer Ollama integration first? And what models are you running locally? Always looking to prioritize based on real usage!
Thanks for the great suggestion!
Olvy
@chrismessina We now have support for local inference in prompt-forge you can now use ollama for local inference thanks to the work by https://github.com/halilugur
This is great! Congratulations on the launch @insaanimanav
Olvy
@nshntarora Thanks
PMFLIX
Olvy
@adhil_sha Thanks Man
Olvy
For everyone checking out the demo - what surprised you most?
And what feature should we prioritize next? Your feedback shapes our roadmap! 🚀
@insaanimanavNice one!! i've upvoted your product and try it out, awesome. We will launch our AI product soon, hope you'll upvote for us too!!
Olvy
@khiemlee Thank you! 🚀 It's been an incredible journey building this. I'll surely checkout what you are releasing . Btw what kind of prompts do you work with most? Code, creative writing, data analysis? Always looking to improve the tool based on real use cases! ✨
@insaanimanav I played with this a little bit the UI is a bit confusing to me. After you generate the evals, I was expecting to be able to save the tests and re-run them but I couldn't figure out how to do that. Is that supported?
Vomyra AI – Voice AI Agent
Olvy
@sumitgoel Love this perspective! You're absolutely right - prompt engineering is becoming a core skill across industries!
Just like how everyone learned Excel or basic coding, systematic prompt engineering will be essential. That's exactly why I wanted to make it accessible with AI assistance and structured testing - democratizing what's currently an art form!
Would love to have you as a user! Try the demo and let me know what you think - especially curious about your experience with the AI-guided prompt generation vs the testing features!
Thank you for the vision and support!
Dirac
That's so cool! I had a comparison with other prompt-engineering tools for my coding. Gotta say, much better result and much faster :)
Olvy
@peter_zhou_overflow That's exactly what I built PromptForge for! What other tools were you testing against? And what made the biggest difference - the AI-assisted generation, the speed, or something else?
Olvy
@raagul_manoharan1 Thank you! And great question!
Yes, I built this myself with AI assistance for parts of the development - felt very meta using AI to build AI prompt engineering tools!
Used Cursor for some of the frontend components, but the core architecture, evaluation algorithms, and product vision were all human-designed. AI was more like a really smart pair programming partner!
The irony wasn't lost on me that I was prompting AI to help build a tool for better prompting!
What's your experience with AI-assisted development? Always curious how other builders are using it in their workflow!
Paraflow
This product really highlights in the built-in analytics and ability to automatically generate test suites. It’s inspiring to bring real software engineering principles to AI prompts.
Olvy
@pulin_yuuu Thank you! You totally nailed the core philosophy - treating prompts with the same rigor we use for code!
The automated test suite generation has been the biggest game-changer. Just like we wouldn't ship code without unit tests, why ship prompts without systematic validation? The analytics help you understand what's actually working vs just guessing!
Are you coming from a software engineering background? Would love to hear your thoughts on other SE principles we could apply to prompt development - version control, CI/CD for prompts, etc.!
Really appreciate the thoughtful feedback!
Olvy
@charlestehio Thanks , agreed on this there is still quite a fair bit of nuance needed to get the AI to do what we want it to