StartupTim

• Anybody using a local model connected to a portable handheld STT/TTS to basically talk to your model? Must be fully local (local models, local whisper/tts/stt).

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 15 comments

• Anybody know of an MCP to speak to chatGPT using your membership to create, view, and edit images?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 1 comments

• I will soon have $100k to build an in-house LLM server. Goal: Best agentic coding model.

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 95 comments

• Best Vibe Code tools (like Cursor) but are free and use your own local LLM?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 98 comments

• Anybody using LMStudio on an AMD Strix 395 AI Max (128GB unified memory)? I keep on getting errors and it always loads to RAM.

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 12 comments

• Best Vibe Code tools that are free and use your own local LLM as of August 2025?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 22 comments

• Any tutorials for using the Nvidia DGX Spark with llama.cpp and models and configuring it?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 3 comments

• Best text-to-image models that support reference images and use openai api standards?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 6 comments

• Any idea how to get ollama to use the igpu on the AMD AI Max+ 395?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 14 comments

• With Unsloth's model's, what do the things like K, K_M, XL, etc mean?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 29 comments

• What's the most affordable cloud-based VM/GPU that works well?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 26 comments

• Postfix server suddenly is sending with relay= field, any idea?

  Posted by StartupTim@reddit | linuxadmin | View on Reddit | 2 comments

• [HELP] Best way to config nginx to handle 25k concurrent users from a former Apache guy

  Posted by StartupTim@reddit | linuxadmin | View on Reddit | 8 comments

• How do I use lemonade/llamacpp with AMD ai mix 395? I must be missing something because surely the github page isn't wrong?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 12 comments

• Tried 10 models, all seem to refuse to write a 10,000 word story. Is there something bad with my prompt? I'm just doing some testing to learn and I can't figure out how to get the LLM to do as I say.

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 98 comments

• Best local coding model w/image support for web development?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 4 comments

• Any idea how to use ollama (debian) with 2x GPUs to load larger models?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 7 comments

• Which - HAPROXY vs Nginx for load balancing websites w/SSL termination on the LB?

  Posted by StartupTim@reddit | linuxadmin | View on Reddit | 3 comments

• What's the current, most affordable cloud GPU option for 16-32ish vram that is on demand for 1-10 minute usages at a time?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 13 comments

• Best coder LLM that has vision model?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 16 comments

• Does certbot/LetsEncrypt with manual DNS challenge require a new TXT record on each renew?

  Posted by StartupTim@reddit | sysadmin | View on Reddit | 17 comments

• Any LLM benchmarks yet for the GMKTek EVO-X2 AMD Ryzen AI Max+ PRO 395?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 15 comments

• Need help with Debian linux Nvidia driver for RTX 5060Ti

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 15 comments

• Any stable drivers for linux (debian) for 5060Ti 16GB?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 12 comments

• Hardware question for general AI/LLM. Would running 2x 5070 Ti 16GB on pcie5 x8 (versus x16) slow things down a lot?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 19 comments

• Any command-line tools to download a huggingface model and convert it to work with ollama?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 13 comments

• What went wrong? Mistral-Small GGUF responds with a huge text about Wordpress when I say "hello".

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 12 comments

• Does (or when) will Openwebui w/ollama API support stable diffusion reasoning models?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 2 comments

• With all the new models dropping recently, which is the best for Python development with a limitation of 20GB VRAM?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 21 comments

• Build request: $2500 "AI in a box" build list request for LLM/SD

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 14 comments

• Got a Mac M4 Mini 24GB for ollama. How do I increase the unified memory limit for the GPU?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 3 comments

• Any instructions for installing ollama as a service on MacOS headless (via SSH)?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 5 comments

• Thoughts on best LLM for simple story writing?

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 4 comments

• OpenWebUI "timing out" issue.

  Posted by StartupTim@reddit | LocalLLaMA | View on Reddit | 2 comments

• Any idea how to stop a Dell Poweredge from attempting NVME on a drive bay that has SaS drives?

  Posted by StartupTim@reddit | sysadmin | View on Reddit | 3 comments