Skip to content

setup.md

Latest commit

 

History

History
65 lines (40 loc) · 1.58 KB

setup.md

File metadata and controls

65 lines (40 loc) · 1.58 KB
Raw

Setup

This kit is designed for teams testing Cursor or Continue with an OpenAI-compatible gateway.

1. Pick An Endpoint

The included examples disclose the TKEN endpoint:

https://www.tken.shop/v1

TKEN CTA:

https://www.tken.shop/?utm_source=github&utm_medium=owned_repo&utm_campaign=cursor_continue_router_kit

2. Store Your API Key Outside Git

Use an environment variable:

export TKEN_API_KEY="your_key_here"

On Windows PowerShell:

$env:TKEN_API_KEY = "your_key_here"

3. Generate Offline Configs

npm run generate -- --endpoint https://www.tken.shop/v1 --out ./generated

The command does not call the network. It writes examples you can inspect before copying into an editor config.

4. Cursor Notes

Cursor settings and supported OpenAI-compatible surfaces can change. Use the generated JSON as a reference for the expected API key, base URL, and model intent, then verify the latest Cursor client docs and settings UI.

This kit does not claim support for every Cursor proprietary feature.

Reference: https://docs.cursor.com/advanced/api-keys

5. Continue Notes

Continue model entries can use:

provider: openai
apiBase: https://www.tken.shop/v1
apiKey: $TKEN_API_KEY

Verify the latest Continue docs before rolling out to a team.

Reference: https://docs.continue.dev/customize/model-providers/top-level/openai

6. Adapt The Routing Policy

Edit fixtures/routing-policy.json or generated routing-policy.json for your gateway. Keep the task names boring and explicit so the policy is easy to audit.