Add installation instructions

README.md

@@ -6,7 +6,25 @@ Maestro is a terminal assistant powered by LLM models, it transforms your instru
 
 ![](maestro.svg)
 
-## Installation
+## Basic Installation
 
-1. Get the latest binary from the [releases page]()
-2. Run it: `./maestro -h`
+1. Get the latest binary from the [releases page](https://github.com/pluja/maestro/releases)
+2. Run it: `./maestro -h`
+
+> You'll have a much better experience if put the binary in a directory that's in your `$PATH` and rename it to `maestro`. For example: `sudo mv ./maestro /usr/local/bin/maestro`. Then you can run it with `maestro` from anywhere!
+
+### Offline Usage with [Ollama](https://ollama.ai)
+
+1. [Install ollama](https://ollama.ai/download), or [run it with docker](https://hub.docker.com/r/ollama/ollama)
+2. Pull some models with `ollama pull <model-name>`
+    - You must pull at least the default `codellama:7b-instruct` model: `ollama pull codellama:7b-instruct`
+3. Start the server with `ollama serve` (default docker behavior)
+4. Set the ollama server URL with `./maestro -set-ollama-url <ollama-url>`
+   - Example: `./maestro -set-ollama-url http://localhost:8080`
+
+### Usage with OpenAI's API
+
+1. Get an API token from [OpenAI](https://platform.openai.com/)
+2. Set the api token with `./maestro -set-openai-token <your-token>`
+
+After setting your token, you can use the `-4` flag to use GPT4-Turbo or `-3` for GPT3.5-Turbo.