Compatibility boundary

Start with Continue Chat, then expand deliberately

Continue’s current configuration format describes a model with a provider, model ID, optional API base, roles, and request options. This tutorial uses that documented OpenAI-provider shape with ScriptEngine’s verified bearer-authenticated /v1/models and /v1/chat/completions endpoints.

The first workflow is intentionally limited to ordinary text chat. Continue can also use edit, apply, autocomplete, context providers, and MCP tools, but those features add request fields or capabilities that must be tested against the exact model. Do not add tool_use or image_input capability flags just because the client supports them.

Continue’s official config.yaml reference is the authority for the current schema. The field names and UI may change between Continue releases, so keep the guide’s date in mind when upgrading.

Requirements

Discover a model before editing config.yaml

Install the current Continue extension or CLI from Continue’s official distribution channel.
Create a private ScriptEngine API key and store it in your secret manager or environment—not in a committed YAML file.
Use the same key to discover a current model ID. The public catalog is helpful for research; the authenticated response is authoritative for the account.

Model discovery

export SCRIPTENGINE_API_KEY="YOUR_PRIVATE_KEY"
curl https://scriptengine.org/v1/models \
  -H "Authorization: Bearer $SCRIPTENGINE_API_KEY"

Copy one text-capable ID from the response and use it in the Continue configuration as MODEL_ID_FROM_DISCOVERY. Never copy a model ID from an old blog post without checking it again.

Configuration

Add a ScriptEngine model to Continue

Continue uses a YAML model entry. The following example keeps the role to chat while you establish connectivity. Continue’s secrets syntax is shown as a placeholder; use the secret mechanism supported by your installed version.

config.yaml

name: ScriptEngine chat
version: 1.0.0
schema: v1

models:
  - name: ScriptEngine model
    provider: openai
    model: MODEL_ID_FROM_DISCOVERY
    apiBase: https://scriptengine.org/v1
    apiKey: ${{ secrets.SCRIPTENGINE_API_KEY }}
    roles:
      - chat

If your Continue build expects a different secret interpolation syntax, replace only the key expression; keep the provider, base URL, and model values unchanged. Do not add capabilities: [tool_use] or capabilities: [image_input] until separately verified.

Verification

Prove the endpoint before opening a repository

Run the API check below before troubleshooting Continue. It distinguishes an invalid key or stale model ID from a local extension configuration problem.

Chat-completions smoke test

curl https://scriptengine.org/v1/chat/completions \
  -H "Authorization: Bearer $SCRIPTENGINE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"MODEL_ID_FROM_DISCOVERY","messages":[{"role":"user","content":"Reply with exactly: Continue connected"}]}'

In Continue, select the configured model and send a short prompt such as “Reply with exactly: Continue connected.” Once that succeeds, test a small, non-sensitive file. Keep autocomplete, edit, apply, Agent mode, and MCP tools disabled until you know the exact request path is supported by the selected model.

Troubleshooting

Resolve common Continue errors

401 or unauthorized: Check that the secret is available to the Continue process and that it is a ScriptEngine API key. Re-run /v1/models with the same environment.
Model not found: Refresh discovery and replace the model field with an ID returned for your key. Provider prefixes and stale IDs from another service are not interchangeable.
Continue calls the wrong path: Use https://scriptengine.org/v1 for apiBase, not the homepage and not /v1/chat/completions. The client appends the route.
Agent or image feature fails: Keep the chat model entry working first. A successful chat response does not prove tool calling, image input, embeddings, or Responses API support.

Cost and safety

Use a predictable budget for coding experiments

Continue can send large file and context payloads. Start with a small prepaid balance, choose a current model after checking the model rates, and monitor usage in the ScriptEngine workspace. The published 86% figure is a dated reference maximum for one comparison, not a guaranteed discount for every Continue request.

Never commit an API key to config.yaml, a repository, a screenshot, or an issue. ScriptEngine currently publishes no uptime or response-time SLA; add client timeouts, error handling, and your own fallback plan for important work.

Sources and next steps

Keep the YAML aligned with Continue’s schema

This page was checked August 4, 2026 against Continue’s current configuration reference and ScriptEngine’s published API contract. Re-check the official reference after a Continue upgrade.

Browse all integration guides Compare an OpenRouter alternative

Use Continue with ScriptEngine.