2226 results found
-
Saving chat convo per task like in cursor ide
Current chat feature in prototyper gets too long and going back to previous request is difficult. It would be nice to create a separate chat convo per task. The prototyper will auto save the task and giving it a title or description about the task. This way it is easier to go back to the task if we want to update or revise related to the task. At the same time easily restore the code to that particular task
2 votes -
attaching a google doc
I wrote my entire req in a google doc and i can't attach it. It's easier to write in a google doc than in the prompt box.
3 votes -
only 5 projects should be able to have 30 projects?
needing to add more projects
2 votes -
for he love of god stop recoding my entire program because i ask a question
days of coding daaaaaaaays lost because i asked a question it recoded the whole thing and didnt even answer the question
2 votes -
6 votes
-
Compress responses served by Cloud CDN
I was running https://pagespeed.web.dev/ and it said I could improve LCP and FCP if I enable text compression. This is a feature of Cloud CDN (https://docs.cloud.google.com/cdn/docs/dynamic-compression), which is the CDN that App Hosting uses, but doesn't seem to be enabled for App Hosting.
1 vote -
Ms pint
Best paint all item tools have in the painter app
1 vote -
Connect
Play store
5 votes -
Master user prompt
allow the user to set a universal context prompt that will always be used in addition to prompts from the chat. It is INFURIATING that it will always describe the plan and implement the changes in a single step. I want it to ALWAYS generate the plan and wait for explicit authorisation to make the changes. It will do that if i ask it to wait in a chat prompt, but i can't make it adopt this behaviour by default. It will always revert back to making changes without authorisation.
I want a master prompt to which i can add…
3 votes -
Ai understanding
I would like to to see you improve the ai understanding of the prompts. And improve the quality of the codes
6 votes -
Rate limiting reads & writes
The Realtime Database should have rate limiting available for the .read & .write properties in the Rules tab.
1 vote -
1 vote
-
The ability to search the web
The ability to search the web to retrieve the latest, version-specific documentation and working code examples directly from official sources. Get accurate, concise, and highly relevant, with no unnecessary content so that its coding capability will be better. Just like 'Context7'
4 votes -
Add C and C++
Please add C and C++
73 votes -
Support for Gradio and Streamlit web applications
Given Project IDX's focus on being integrated with generative ai — project templates for Gradio and Streamlit to demo ML/Gen AI systems would be a great addition to the currently available templates. Both as standalone templates and frameworks for Google's Gemini API template.
This can also be a great opportunity to showcase Google's generative ai models via Gradio/Streamlit.
Note: My idea is intended to use Python as the primary language (specific to Gradio since there's also a JS (
@gradio/atoms)). Though the potential for polyglot Gradio and Streamlit projects should also be supported, respective to the framework. For example,…7 votes -
Show unsplit graph on crash details page
The crash details page crash chart doesn't tell me overall crash performance. There is no feature that allows me to see the chart as a whole to determine if overall crashes have gone down.
I use the crash list page to see the chart as a whole because it's easier to visually determine if the crash rate has gone down across versions that way.
1 vote -
1 vote
-
Put a timestamp on Prompts and Response for easy tracing.
It would be nice to put a timestamp on all our chats and responses so that whenever one wants to role back, it will be easy to trace which response was last published. If this can be added as a feature it'd be nice.
2 votes -
Multi device collaboration
Currently people cannot work on separate branches simultaneously. Only one branch is active even if multiple people are using the project in the studio. It would be a good feature to have where different people can work on different branches at a time. with each branch being locked to a specific user and other user cannot changes until that branch is released. Team collaboration will be much more effecient and people can work on different features at the same time.
3 votes -
Way to handle doom loop coding
I've been running into an issue repeatedly letting Firebase Studio AI try to fix an error that it keeps re-detecting. It acknowledges that it's a persistent issue, but none of the fixes seem to work. I'm not sure how to fix it either. Maybe it would be good to have a way for the AI at some point engage human support on the Google side? Or maybe open up for a request to identify and investigate problems the AI seems to be struggling with?
Here's the particular error in this instance:"Runtime Error
Error: ./nodemodules/@opentelemetry/sdk-trace-node/nodemodules/@opentelemetry/context-async-hooks/build/src/AsyncHooksContextManager.js:20:20
Module not found:…4 votes
- Don't see your idea?