Continue with Verdictan

Add your Verdictan gateway as a model in Continue's config.yaml, and Continue's requests for that model go through your policies. The same configuration works in VS Code and in JetBrains IDEs. Start with chat, then turn on edit and apply one at a time.

Before you start

• A gateway on http://127.0.0.1:41002 with a provider target for the model you plan to use. See Gateway Setup for IDEs.
• A client API token. See API Tokens for IDEs.
• The Continue extension in VS Code or your JetBrains IDE.

Configure Continue

1. Put the token in ~/.continue/.env, outside your project:

   ~/.continue/.env

   VERDICTAN_CLIENT_TOKEN=<client-api-token>

   If you use an environment file inside the project instead, add it to .gitignore.

2. From the Continue panel, open the active local configuration and add this model, keeping your other entries:

   config.yaml

   name: Verdictan
   version: 1.0.0
   schema: v1
   
   models:
     - name: Verdictan Chat
       provider: openai
       model: your-model-id
       apiBase: http://127.0.0.1:41002/v1
       apiKey: ${{ secrets.VERDICTAN_CLIENT_TOKEN }}
       roles:
         - chat

3. Replace your-model-id with an id from the gateway's authenticated GET /v1/models response.

4. Restart the IDE so Continue reads the secret, then pick Verdictan Chat in the model picker.

Once chat works, add edit and apply to roles and test each one. Add autocomplete only if the upstream model supports fill-in-the-middle. See the Continue config reference and how Continue resolves secrets.

Verify

1. Check the token and the model list:

   export VERDICTAN_CLIENT_TOKEN="<client-api-token>"
   
   curl -fsS \
     -H "Authorization: Bearer $VERDICTAN_CLIENT_TOKEN" \
     http://127.0.0.1:41002/v1/models

2. In Continue, pick Verdictan Chat, send Reply with Hello., and check that Continue shows the answer.

3. In a second shell, read the recent events:

   verdictan events tail --since 10m --json

4. Check that the newest event has a timestamp from the moment you sent the prompt. Delivery can take a few seconds, so run the command again if the event is not there yet.

If no event appears, work through IDE Integration Troubleshooting in order.

Troubleshooting

Symptom	Cause	Fix
Connection refused	The gateway is not running on port 41002	Start it, and check /healthz with curl.
401 Unauthorized	Continue did not resolve the secret, or the token is inactive	Check the .env location and variable name, restart the IDE, and check the token in the Console.
Continue rejects the model	The ID does not match a provider target	Use an id from the authenticated GET /v1/models response.
A reply but no event	Continue used another model	Pick Verdictan Chat and check the active config.yaml. Don't use VS Code's proxy setting.
Autocomplete fails	The model doesn't support fill-in-the-middle	Remove the autocomplete role, or use a model that supports it.

Next steps

• Configure Policies for IDE Traffic: controls for the prompts Continue sends.
• Monitor IDE AI Usage: Continue traffic in events, History, and Usage.