Tool setup
Continue.dev + SeaLink
Connect Continue.dev to SeaLink with the current config.yaml v1 format. One configuration works across CLI, VS Code, and JetBrains.
Prerequisites
- VS Code or JetBrains IDE
- Continue extension installed
- A valid SeaLink API Key
Setup Steps
Install Continue
CLI: run npm install -g @continuedev/cli in your terminal. VS Code: open the Extensions panel (Ctrl+Shift+X), search for 'Continue', and install. JetBrains: go to Settings → Plugins, search for 'Continue', and install. Pick the option that matches your workflow — one install covers all your needs.
Open the Continue config file
Continue stores its configuration at ~/.continue/config.yaml. You can open it with any text editor. If you're migrating from an older config.json, we recommend keeping only config.yaml as your active config file. Having both schemas active at once can lead to unexpected behavior.
Add SeaLink models to your config
Add the following to your config.yaml: name: SeaLink version: 1.0.0 schema: v1 models: - name: SeaLink GPT-4o Mini provider: openai model: gpt-4o-mini apiBase: https://test.sealink.io/v1 apiKey: sk-sealink-your-key roles: - chat - edit - apply - autocomplete Want multiple models? Just add more entries under models — one per model. For example, you could add a Qwen model for chat and a Claude model for code editing, all in the same config.
Save and verify
Once you save config.yaml, Continue picks up the changes automatically — no restart needed. For CLI users: run continue --help to confirm the install, then send a simple prompt to test connectivity. In the IDE: open the Continue chat panel and send a message. If you get a response, and autocomplete suggestions appear as you type code, you're all set — SeaLink is connected.
Common Issues
Config file fails to parse
Double-check that your config.yaml has the required top-level fields: name, version, and schema: v1. Also make sure every model entry includes name, provider, model, apiBase, and apiKey. Missing any one of these will cause a parse failure — compare against the example above to be sure.
Chat works but autocomplete never appears
This usually means the model's roles list is missing 'autocomplete'. In the current v1 schema, all model capabilities are declared through the roles field — there's no separate tabAutocompleteModel config. Check that your model entry includes autocomplete in its roles.
Model returns a 401 authentication error
A 401 almost always points to an API Key or Base URL mismatch. Check that apiBase ends with /v1 and that apiKey starts with sk-sealink- and is the full key string. Try copying the key again from your SeaLink dashboard to rule out copy-paste issues.
How do I use Anthropic models (like Claude) in Continue?
Continue supports the Anthropic provider natively. Set provider to "anthropic", apiBase to https://test.sealink.io/anthropic, and model to your desired Claude model ID (e.g. claude-sonnet-4-6). Everything else stays the same.