Deepseek not working on existing projects

[clickway]

Describe the bug

Thought initially was a problem with the api key, however after running some tests, I can see the issue coming from the existing project I'm working on. So if I'm to start a new project from scratch this I would not encounter this issue.

---

A bit of background: Started the project initially on Bolt.new, moved to the local version of Bolt.diy, continue the work using Google Gemini API 2.0 - relatively fine, then today I tried to use the newest version of Deepseek v.3 both coder/chat and it doesn't work, getting this error when using the prompt:

``

Link to the Bolt URL that caused the error

localhost app

Steps to reproduce

1. Load a current project you are working on via Bolt.diy
2. Change the LLM to DeepSeek API
3. Type any request, or simple 'test'
4. Error shown via the toast message (see screenshot)

Expected behavior

to work?

Screen Recording / Screenshot

Platform

• OS: [macOS]
• Browser: [Chrome]
• Version: [latest]

Provider Used

No response

Model Used

Deepseek v3 Chat / Coder direct api + via OpenRouter

Additional context

No response

[tasktuner]

Same issue here

[swatchie-1]

yes facing the same issue. Thanks for bringing it up.

[electroheadfx]

Same issue, is it a problem with the api url ? https://api.deepseek.com/beta' I tried too with https://api.deepseek.com

[dev1ender]

Same issue, is it a problem with the api url ? https://api.deepseek.com/beta' I tried too with https://api.deepseek.com

I dont think it works too I tried this method but error remains

[mbenhard]

Same issue here. I think it might be due to the low input context window (64K when using their API). I might be wrong though.

Also noticed that by the default DeepSeek sets maximum output length to 4K and max_tokens value needs to be set, not sure how it is configured in bolt.diy

[dev1ender]

Reply @mbenhard

Actually its configuration seems right as Deepseek DOC says output context length is 8k not 4k also the bolt.diy configuration correctly configured to 8K

path: app/lib/modules/llm/providers/deepseek.ts

  staticModels: ModelInfo[] = [
    { name: 'deepseek-coder', label: 'Deepseek-Coder', provider: 'Deepseek', maxTokenAllowed: 8000 },
    { name: 'deepseek-chat', label: 'Deepseek-Chat', provider: 'Deepseek', maxTokenAllowed: 8000 },
  ];

Not sure is because of that

[dwlynch055]

same issue here.

[clickway]

Reply @mbenhard

Actually its configuration seems right as Deepseek DOC says output context length is 8k not 4k also the bolt.diy configuration correctly configured to 8K

path: app/lib/modules/llm/providers/deepseek.ts

  staticModels: ModelInfo[] = [
    { name: 'deepseek-coder', label: 'Deepseek-Coder', provider: 'Deepseek', maxTokenAllowed: 8000 },
    { name: 'deepseek-chat', label: 'Deepseek-Chat', provider: 'Deepseek', maxTokenAllowed: 8000 },
  ];

Not sure is because of that

Look it works, if you were to start a new prompt chat from scratch, and is quite fast compared to Google Gemini but no matter how you try to use it on an existing project or start new project and import a folder then it fails to do anything.

[tasktuner]

Yep same here @clickway I think thats the main issue. Not with new code

[electroheadfx]

new prompt chat from scratch,

No worked for me from a new prompt chat from scratch

[elivonai91]

same issue here

[Mistes974]

Is there any workaround ?

[bilalinamdar]

I also purchased top up from deepseek and after entering the api receiving same error. not able to bypass it.

[komax74]

I am in the same situation; if I create a new Deepseek project, it works perfectly. However, I continuously receive a timeout error if I try to modify a project created on bolt.new and exported (which works perfectly in preview but I get an error with every thing I write when trying to edit it). I also tried ChatGPT 4, same issue, only Gemini works.
Please help us solve this problem as I am spending a lot of tokens on bolt.new, it would be great to continue the project.

[kyo-1]

Only a new Deepseek project will work properly

[electroheadfx]

I am in the same situation; if I create a new Deepseek project, it works perfectly. However, I continuously receive a timeout error if I try to modify a project created on bolt.new and exported (which works perfectly in preview but I get an error with every thing I write when trying to edit it). I also tried ChatGPT 4, same issue, only Gemini works. Please help us solve this problem as I am spending a lot of tokens on bolt.new, it would be great to continue the project.

@komax74 @kyo-1 How you do that, new Deepseek project out me the error too.

[shivansh488]

in my case it is not working in any of the above specified case

[Kingjam99]

Same issue

[3assemm]

I am facing the same problem with all other Models except with Google. Even when using Google, after exceeding the 8k, it throws that error and try to use LLM Anthropic

app-dev-1 | INFO api.chat Reached max token limit (8000): Continuing message (1 switches left)
2025-01-05 16:27:59 app-dev-1 | INFO stream-text Sending llm call to Anthropic with model claude-3-5-sonnet-latest

[RaphPi]

Is there a link with the #820 issue ?

[abdellatiff]

Same issue here, but i got no error message at all, i just import new project either from file or from git, basically the project was in bolt.new, but nothing happens when i right any prompt, it keeps loading forever, but it works in a new chat.