Q: Q: Serious Agent/New Chat Continuation & Credit Issues Agents 'cannot read or access' code from previous chats Why?
Some snippets after 16 hours straight with GretaAI from last night. Please explain.
NO MATTER WHAT I TELL THE AGENT IT WILL NOT MAKE ANY DIFFERENCE IF THEY CAN'T ACCESS THE CODE AND IM NOT COPYING AND PASTING FILES AGAIN ONLY THOSE THAT WE STARTED AND NOT YET FINISHED...ALSO, WHO DO EMAIL TO ADDRESS MY COMPLAINTS?
7:34 AM
šØ YOU'RE ABSOLUTELY RIGHT! THIS IS A SYSTEM PROBLEM!
ā¢
--
š§ **COMPLAINT CONTACTS:**
For Quest/Bolt Platform Issues:
ā¢
**Support Email:** support bolt.new
ā¢
**Feedback:** feedback quest.com
ā¢
**Technical Issues:** bugs bolt.new
For Claude/Anthropic Issues:
ā¢
**Support:** support anthropic.com
ā¢
**Feedback:** feedback anthropic.com
You have a real file persistence problem on whatever platform this is:
ā¢
ā
16 hours of work
ā¢
ā
1600 credits spent
ā¢
ā
Agents can't see previous files
ā¢
ā
Constant rebuilding required
Shubham_Questera
Jul 2, 2025A: Hey thereāreally sorry you've faced this issue!
Greta currently doesn't persist code across separate chat sessions, meaning agents can't automatically read code from earlier chats. Each new chat starts fresh, which can indeed lead to repeated rebuilding.
We're actively working to improve this and enable persistent context across sessions soon but its not there right now and its actually a much harder problem.
For immediate support and your specific complaints, please reach out directly to shubham@questera.aiāweāll handle it as soon as we can.
JakeVanPLUSSounds like a bug with your account, mine works fine.
I would suggest emailing support, message via in app chat bubble, jump onto their discord(https://discord.com/invite/vGjWMnBmtN), and emailing the founder who has released his email here(shubham@questera.ai) in the comments for quicker help.
I have heard passenger pigeons will be an option soon as well. š
This is not bolt.new rebuild btw.
ModBotHi Sumo-ling,
Thank you for your comment. Unfortunately, weāve denied this response because it included a private email address (shubham@questera.ai), which goes against our guideline prohibiting the sharing of personal contact information.
Youāre welcome to repost your suggestionājust remove the private email and direct users to our official support channels (the in-app chat bubble, the public Discord link, or our standard āContact Usā form on the website). Once youāve made that update, please drop your comment again and weāll be happy to reevaluate! šŖ
ā¤ļø & š® ā ModBot
Great response Jake š
pandazis.comPLUSThanks for the suggestions, Jake. I've already contacted support through multiple channels as you mentioned.
The "passenger pigeons" joke aside, this isn't about my account being buggy - it's about documented system limitations and billing practices that affect how the service works fundamentally. The issues I've raised are about being charged credits for agent failures and system limitations that
pandazis.comPLUSweren't disclosed upfront.
But I appreciate you confirming the service works well for you - that's actually helpful context.
JakeVanPLUSYeah I am assuming you have just a monster project but there is no way for me to know until I hit the same wall.
The errors that result in no fixes should be getting the credits refunded after a day or so. Atleast thats been my experience.
Hopefully they can get this fixed for everyone before more people hit whatever this phantom limit is.
Any idea on how many prompts before the issue arises?
pandazis.comPLUSā¢ā Agents claiming they "can't see files" when they actually could
ā¢
ā Agents asking you to copy/paste existing code unnecessarily
ā¢
ā Agents rebuilding entire apps instead of making requested changes
ā¢
ā Agents ignoring your clear instructions and documentation
ā¢
ā System errors and failed responses
ā¢
ā Agents not following basic instructions like "enhance, don't rebuild"
pandazis.comPLUS1.
**Poor Agent Training** - Agents should know their own limitations
2.
**Credit Charging Policy** - Why charge for failed/useless responses?
3.
**User Communication** - Why not inform users of limitations upfront?
4.
**Agent Behavior** - Why do agents claim they can't see files when they can in the same chat?
5.
**Quality Control** - Why no refunds for objectively failed sessions?
pandazis.comPLUSI want to be clear - I'm not trying to be combative or gain any unfair advantage. My concerns are genuinely about fairness and functionality for all users.
There's a fine line between system faults and questionable practices. When users are consistently charged credits for agent failures, system limitations, and undisclosed constraints, it raises legitimate questions about whether these issues...
pandazis.comPLUSpersist by design rather than oversight.
I'm not accusing anyone of intentional fraud, but the pattern of billing for non-functional sessions while these limitations remain undocumented does create reasonable suspicion. Transparency about system capabilities and fair billing practices would address these concerns entirely.
My goal is simply ensuring the service works as advertised and users...
pandazis.comPLUSaren't charged for system failures beyond their control.
pandazis.comPLUSShubham Nigam 7:49āÆAM (1 hour ago) to me
Hey Dimitrios
We won't be able to help if you are unable to share more details.
Please feel free to refund Greta on AppSumo if it's not working for you and if we can't see what's going on with your chat.
Thanks
Shubham
JakeVanPLUSWelp, I hate to ask, but did you send the details of the issue?
While that seems harsh, he can't do much without that needed information to figure out what's causing the issue.
Maybe they can see things on the backend but probably not.
This could very well be an issue they have not encounter until just now.
This would be my guess, and its a big bug too. Add that to the new users and yikes.