How to use the AshnaAI API in VS Code

Watch the VS Code setup

The clip below is the Language Models overlay in VS Code: Add Models is the control that starts a Custom Endpoint. Someone who already has a key can follow the clicks, then return here for the exact URL and model id.

Step 1: Create an AshnaAI API key

The editor needs a Bearer token before it can list or call catalog models. Open Account → API. Create key asks for a name—VS Code is enough. The secret is shown once. It is sent as Authorization: Bearer. It does not belong in git, a gist, or a recorded screen.

Step 2: Open Language Models in VS Code

In the Chat view, open the model picker and choose Manage Language Models. The Command Palette command is Chat: Manage Language Models. The Language Models editor opens as a modal over the editor—the same overlay as the video. Official notes: AI language models in VS Code.

Step 3: Add a Custom Endpoint

Select Add Models, then Custom Endpoint. That provider is the OpenAI-compatible path. Group name: AshnaAI. API type: chat-completions. Base URL: https://api.ashna.ai/v1/api Paste the key into the secret field VS Code stores locally.

When the provider URL is https://api.ashna.ai/v1/api model discovery uses GET /v1/api/models—the same list as How to call any catalog model through the API. A model can also be written by hand in chatLanguageModels.json with vendor customendpoint, id glm-5.3-flash, and url https://api.ashna.ai/v1/api/chat/completions

Step 4: Pick a catalog id

After the provider saves, the Chat model picker lists the catalog rows VS Code discovered. glm-5.3-flash is the usual coding pin. Other ids work the same way: change only the model string. Confirm ids on model and agent ids or list models.

Tool-calling models stay available for agent chat. If a newly added row does not appear, Microsoft’s docs say restart VS Code once.

Fields that have to match

These are the values the Custom Endpoint must send. A typo on the base URL is the usual failure: https://api.ashna.ai without /v1/api points at the wrong path.

Try it now

Create a key in Account → API, then add the Custom Endpoint in VS Code. HTTP examples live on AshnaAI API docs and chat completions. To try Flash in the product UI first, open GLM-5.3-Flash in chat.