Local AI Models

Run RespectASO Pro's AI features on a model hosted on your own Mac — instead of a cloud provider. The AI Niche Researcher, AI Competitor Analyzer, and ASO Score Simulator all work exactly the same, but the AI runs on-device.

Why run AI locally?

  • No per-query cost. No API key, no metering — run as many analyses as you want.
  • Total privacy. Your prompts never leave your Mac.
  • Works offline. No internet round-trip for AI runs.
  • Your model, your choice. Pick any local model and switch freely.

What you need

Local AI is part of RespectASO Pro and runs in the native macOS app. You'll also need a local model server — RespectASO works with two popular, free options:

  • Ollama — a simple command-line model server (ollama.com)
  • LM Studio — a friendly desktop app for running models (lmstudio.ai)

Any other OpenAI-compatible local endpoint works too, via the Custom option.

Set up with Ollama

  1. Install Ollama from ollama.com and open it
  2. Download a model, e.g. run ollama pull <model> in Terminal (larger models give more robust results — see Performance below)
  3. In RespectASO, go to Settings → AI and choose Local AI
  4. Pick the Ollama runtime (the address defaults to http://localhost:11434)
  5. Click Detect to list the models you've installed, and select one
  6. Click Test Local AI to confirm your model is ready, then save

Set up with LM Studio

  1. Install LM Studio from lmstudio.ai and open it
  2. Download a model from inside LM Studio, then start its local server
  3. In RespectASO, go to Settings → AI and choose Local AI
  4. Pick the LM Studio runtime (the address defaults to http://localhost:1234)
  5. Click Detect to list available models, and select one
  6. Click Test Local AI to confirm your model is ready, then save

Test before you run. The one-click Test Local AI check runs a quick, representative task so you know your setup is ready before you start a full analysis.

Performance & your hardware

RespectASO's analyses are agentic — they research, score, and refine over several steps — so they involve demanding processing. Local models won't perform equally on every device: speed depends on your Mac's hardware (chip and, importantly, memory bandwidth). If a run feels slow or its suggestions feel thin, a larger model is well worth it — bigger models handle the multi-step workflows best and return more robust, more relevant suggestions.

You can cancel a run at any time, and you can always switch back to a cloud provider for a heavier job and return to Local AI for everyday work.

Privacy

With Local AI, the entire AI workflow stays on your machine — your prompts and results never leave your Mac, and there's no API key to manage. It's the same privacy-first principle as the rest of RespectASO, now extended to the AI features themselves.

Prefer the cloud?

Local AI is one option among several. You can also use your own OpenAI, Anthropic, or Google Gemini key, and switch between any of them anytime from Settings → AI. Learn more about what each tool does in AI Features.