Freezing on every prompt in the chat

Every time I give an indication in the chat, everything freezes and I have to wait about 8 minutes until it comes back and Every time I give an instruction in the chat, everything freezes and I have to wait about 8 minutes until it comes back and shows all the changes and results.

How can I fix it?

I use a Mac Book Pro with Google Chrome installed in Docker on a server with sufficient resources.I use a Mac Book Pro with Google Chrome installed in Docker on a server with sufficient resources.

I also removed in C settingsI also set the Chrome security settings to not pause due to lack of resources.

How can I solve it?

Are you using a local model? Did you try with deepseek or others?

Model usage via API. Grok Coder and Qwen3 Code via Model usage via API. Grok Coder and Qwen3 Code via OpenRouter.

My VPS has 16GB of RAM and is controlled with htop, not even reaching 7GB peak.

can you test this pr : perf: optimize generation response time with stream buffering and caching by embire2 · Pull Request #1985 · stackblitz-labs/bolt.diy · GitHub

Of course, I’ll try it out. Anything I can test and contribute, I’ll be happy to share. I love the project and your dedication. I’m not a developer, but I can help with whatever you need.

I’ll try it and let you know.

How can I add that branch? I use Dokploy.

This way it doesn’t download:

services:
  app:
    #image: ghcr.io/stackblitz-labs/bolt.diy:latest
    image: ghcr.io/stackblitz-labs/bolt.diy:pr-1985
    pull_policy: always
    environment:
      - NODE_ENV=production
      - COMPOSE_PROFILES=production
      - PORT=5173
      - GROQ_API_KEY=${GROQ_API_KEY}
      - HuggingFace_API_KEY=${HuggingFace_API_KEY}
      - OPENAI_API_KEY=${OPENAI_API_KEY}
      - ANTHROPIC_API_KEY=${ANTHROPIC_API_KEY}
      - OPEN_ROUTER_API_KEY=${OPEN_ROUTER_API_KEY}
      - GOOGLE_GENERATIVE_AI_API_KEY=${GOOGLE_GENERATIVE_AI_API_KEY}
      - OLLAMA_API_BASE_URL=${OLLAMA_API_BASE_URL}
      - XAI_API_KEY=${XAI_API_KEY}
      - TOGETHER_API_KEY=${TOGETHER_API_KEY}
      - TOGETHER_API_BASE_URL=${TOGETHER_API_BASE_URL}
      - AWS_BEDROCK_CONFIG=${AWS_BEDROCK_CONFIG}
      - VITE_LOG_LEVEL=${VITE_LOG_LEVEL:-info}
      - DEFAULT_NUM_CTX=${DEFAULT_NUM_CTX:-32768}
      - RUNNING_IN_DOCKER=false
    extra_hosts:
      - "host.docker.internal:host-gateway"
#    command: pnpm run dockerstart
    command: pnpm run dockerstart