Portkey's prompt engineering platform lets teams build, test, and deploy AI prompts at scale. Includes prompt playground, version control, collaborative editing, and AI gateway integration supporting 1600+ models and AI agent frameworks.
Free
Launch Team / Built With
CloudflareYour next app belongs on Cloudflare Workers. Deploy for free
Promoted
@ayush_garg_xyz Portkey’s Prompt Studio solves so many pain points like side-by-side model comparison, version control, and AI-powered optimization. Congrats on the launch
@ayush_garg_xyz Do you support APIs for data import as part of the prompt eval process?
@ayush_garg_xyz Finally, a proper toolkit for prompt engineers! Working with AI models has always felt a bit like trial and error, so having a dedicated studio with version control and real testing tools is a game-changer. Love the focus on collaboration too
@ayush_garg_xyz @jontronic Hey Jonathan, no that is not supported at the moment - we are building a better Evals feature on top of this now.
@ayush_garg_xyz @anas_turki Thank you Anas!
This is an amazing idea, i've been looking for a platform to help me with creating the prompts. I hope this would be it. Giving it a go today!
congratulations on the launch and keep it up!
@ramywafaa Glad you like this Ramy, thank you!
Hi @ramywafaa thanks for the excitement! 🎉 we’re thrilled you’re giving it a go. Hope it makes your prompt creation process easier and faster—looking forward to hearing how it works for you!
This looks amazing @vrv18 and @ayush_garg_xyz ! Love how Portkey streamlines prompt engineering with powerful comparisons and collaboration.
How does your version control system handle prompt iterations? Can teams see performance changes over time and easily roll back to the best-performing versions?
@ayush_garg_xyz @harkirat_singh3777 Thank you Harkirat! We don't have a polished evals product right now - you can run tests on prompts with our API but not in the UI. However, of course all prompt iterations are versioned and you can rollback, use any of the versions easily.
Is this more for system prompts or can be used for general users so they know how to ask better questions?
hey @lylia_djaitpaulien great question! you can play with both system prompt as well as user prompts in portkey playground. you can also use prompt variables inside your prompt to get your user's context
@ayush_garg_xyz @amanintech Thanks so much Aman for your comment. We are also a fan of what you're building at Lamatic!
Onto your suggestions:
1 - This is a must have, and is coming soon to Portkey
2 - Thi s is already supported! Check out https://portkey.ai/docs/integrations/llms/byollm and https://portkey.ai/docs/api-reference/inference-api/gateway-for-other-apis
3 - This is supported on app.portkey.ai as well
4 - We have released an update for this as well, and you can create multiple Portkey workspaces inside a single org.
Been using Portkey for a coupe of years now and the new prompt engineering studio has been a perfect match to write and iterate prompts at scale. Excited for the launch, lightspeed!
@prrranavv Thank you so much Pranav for your feedback and suggestions so far! They have been instrumental in building the Prompt Studio.
Amazing product! This addresses a critical pain point in the AI workflow! The version control with labeled deployments is what's missing in most prompt engineering processes. As someone who constantly tweaks prompts for different use cases across product, business, and content, the ability to compare models side-by-side would save countless hours of guesswork.
Have you considered adding a prompt performance analytics feature that tracks which versions perform best across different contexts or user segments? That kind of data-driven insight could help teams standardize on proven patterns rather than relying on individual intuition.