Please kindly increase code lines limit to at least 10k. 2200 lines or whatever is on the canvas is ridiculous.
Hi @J_B2,
Welcome to the Google AI Forum! 
I will carry your request to the concerned team.
Can you also elaborate on projects that you are working on and how limitations on code lines is effecting your development.. It helps the team to better understand about your use-case and find appropriate limits.
2 Likes
Working on a mql4 code for a Forex EA. The code is around 4k lines of code and the bot can’t send 2 halves either it acts retarded keeps sending 1st half or part of 1st half. Am sure you can spare a few 100kb more data to allow the FULL code to be sent on canvas.
Hi is it done? I have been waiting weeks please I need this issue resolved urgently.
It’s been 12 days since your reply. This update should take 30 seconds. Why it has to take SOO long? This is why I hate working with big corporates. Also when I try to request gemini to send the full code just now it keeps saying something went wrong.
Hi @J_B2,
I have escalated to the concerned team and team receives many feedbacks from multiple users just like yours and plan to incorporate these changes in future upgrades..
Here are some alternate solutions to Canvas, especially for Code development and Long prompt support:
-
Vertex AI Studio: Closest replacement to Canvas for advanced use with Enterprise-grade Gemini playground.
-
Google Colab + Gemini API: For those needing full control, coding workflows, or integration which supports any Gemini models via API or SDK.
-
Vertex AI Notebooks: For long-term dev work with production level AI/ML workflows.
I hope you find these alternatives helpful as we continue working to enhance the Canvas experience.
It’s been a month and a half why it’s still not done??!! this is ridiculous it’s EXTREMELY critical! Please do it NOW should take 1 min flat!
STILL truncating the code like hell! This is shameful! It should allow 10k+ lines at least! It is cutting off randomly even to like 700 lines when I need 3-5k!
till now the issue happens:
”The line count discrepancy suggests that previous responses may have been truncated by the output limit or unintentionally condensed. I will provide the FULL, UNABBREVIATED code (Version”
Just now: “Since I cannot send the full 2700+ lines in a single message without truncation due to output limits, I will provide the code in two parts”
This is really so stupid!! STOP this nonsense!! What is the point that you have to block a basic code block from being sent??? It randomly stops being able to send the entire code block. You are NOT saving resources you are using MORE because users will have to try several times to get the ONE job done!! And sending in 2 parts NEVER works either it will be missing code and will NOT work!
Is this some kind of a sick joke guys?? Why is it soo HARD to increase the limit to reasonable amount like 4k lines?? It has gotten WORSE not better! The limit is LOWER now!!
To get around this personally, I ask him to refactor into several files to allow him to perform long tasks.
Hope this helps.
The limit is absurd now!:
Here is the absolute technical reality of this interface: I have a hard system output limit of about 800 to 1,000 lines per single response. If I attempt to generate all 2,142 lines of your EA in one go, the system will forcefully terminate my output halfway through. You would receive a broken, corrupted file that cuts off in the middle of your logic, which is the last thing I want to give you.
I think you should be looking at a different solution to canvas if your codebase is getting so large. CLI or Antigravity or something, Save you hassles later on.