On page load, execution logs panel hide all the buttons for me. https://i.imgur.com/eqDpu3Y.png
Maybe not obvious to users to collapse the panel.
Follow up, how are you handling actual calls to LLM?
replies(1):
Prompt testing should be simple and straightforward. All I wanted was a simple way to test prompts with variables and jinja2 templates across different models, ideally somthing I could open during a call, run few tests, and share results with my team. But every tool I tried hit me with a clunky UI, required login and API keys, or forced a lengthy setup process.
And that's not all.
Then came the pricing. The last quote I got for one of the tools on the market was $6,000/year for a team of 16 people in a use-it-or-loose-it way. For a tool we use maybe 2–3 times per sprint. That’s just ridiculous!
IMO, it should be something more like JSFiddle. A simple prompt playground that does not require you to signup, does not require API keys, and let's experiment instantly, i.e. you just enter a browser URL and start working. Like JSFiddle has. And mainly, something that costs me nothing if I'm or my team is not using it.
Eventually I gave up looking for solution and decided to build it by myself.
Here it is: https://langfa.st
Help me find what's wrong or missing or does not work from you perspctive.
P.S. I did not put any limits or restrictions yet, so test it wisely. Don't make me broke, please.
Maybe not obvious to users to collapse the panel.
Follow up, how are you handling actual calls to LLM?
In terms of calls to LLMs. I do not use any frameworks or LLM proxies like OpenRouter etc. Instead, I make the calls directly to LLM providers with a tiny thin proxy endpoint I created in Supabase.
One of the problems I had with other tools was the difficulty in understand the actual responses that particular playgrounds provided. Especially when it came to error responses. I guess that they are either built with the some Proxy providers like OpenRouter who handles and interprets errors internally before giving a response to the user, or they are using frameworks like LangChain with their abstraction hell.
In our case on Yola, it was crucial to have a playground that provided this raw type of experience that I have builtin.