Skip to content

getting-started.md

Latest commit

 

History

History
34 lines (20 loc) · 1.67 KB

getting-started.md

File metadata and controls

34 lines (20 loc) · 1.67 KB
Raw

Getting started

Русская версия

Connect to an LLM

The main option is local llama.cpp with an OpenAI-compatible API. The same protocol works with cloud endpoints.

In General:

  1. Base URL - for llama.cpp the default is http://127.0.0.1:8080. For the cloud, use the provider URL (for example https://api.openai.com).
  2. API key - stored in VSCode SecretStorage (not in settings.json). For local llama.cpp it is usually not needed - leave it empty.
  3. Model - pick from the list (loaded from the URL) or type it manually (model name/alias on the server).
  4. If needed, set the key header and scheme (defaults: Authorization + Bearer).

Save the settings. If URL/model are missing, chat shows a hint and a button to open settings.

First request

  1. Ask mode - normal chat; selected editor code is added to context.
  2. Agent mode - the model can call tools (read/edit files, and so on).

Switch mode with the select next to the input (or slash /ask / /agent / ...); default is Settings -> Chat.

Optional project scaffold

Gen: Initialize project (.gen) or chat /init creates .gen/ (agents/commands/skills, ...), seeds .genignore for path/folder ignore, and a starter .gen/config.json for command/redact policy. Not required to chat or index - see security.md and codebase-index.md.

Check

  • In Ask, ask something simple - you should get a streaming reply (if the server supports stream).
  • In Agent, ask to “read README” - a read_file tool-call should appear.

Next: Chat and agent, Settings, Security.