Config ⚙️

aj reads a YAML configuration file from your platform directory.

📍 Paths

• macOS: ~/Library/Application Support/com.awful-sec.aj/config.yaml
• Linux: ~/.config/aj/config.yaml
• Windows: C:\\Users\\YOU\\AppData\\Roaming\\awful-sec\\aj\\config.yaml

🧾 Example

api_base: "http://localhost:1234/v1"
api_key: "CHANGEME"
model: "jade_qwen3_4b_mlx"
context_max_tokens: 8192
assistant_minimum_context_tokens: 2048
stop_words:
  - "<|im_end|>\\n<|im_start|>"
  - "<|im_start|>\n"
session_db_url: "/Users/you/Library/Application Support/com.awful-sec.aj/aj.db"
should_stream: true

🔑 Key Fields

• api_base: Where requests go.
• api_key: Secret authorization key (optional).
• model: LLM model to use.
• context_max_tokens: Context length to request from model.
• assistant_minimum_context_tokens: Will always eject messages from the context if needed to make room for this many tokens to fit the response.
• stop_words: Tokens to cut off model output.
• session_db_url: SQLite DB path for sessions (optional). = should_stream: Whether to stream the response or return it all at once when the inference ends.

✍️ Editing Tips

• After edits, re‑run your command (no daemon reloads required).
• Make sure you include the port number your LLM inference server is running on.
• If your using an online service you can usually create an API key on your account settings or api page.
• jade_qwen3_4b_mlx is a highly capable Qwen 3 4B model that I finetuned for Apple focused systems programming and general discussion. You can find it on huggingface or download it directly in LM Studio.