Continue.dev Not Working? Fix Model Config, Autocomplete & VS Code Errors

config.json not configured — getting started

Continue.dev requires a ~/.continue/config.json with at least one model provider. Open it from VS Code with Ctrl+Shift+P → "Continue: Open Config". A minimal working configuration:

{
  "models": [
    {
      "title": "Claude 3.5 Sonnet",
      "provider": "anthropic",
      "model": "claude-3-5-sonnet-20241022",
      "apiKey": "sk-ant-..."
    }
  ],
  "tabAutocompleteModel": {
    "title": "Starcoder2 3B (local)",
    "provider": "ollama",
    "model": "starcoder2:3b"
  }
}

Config location: ~/.continue/config.json (Linux/Mac) or %USERPROFILE%\.continue\config.json (Windows).

Ollama connection refused

Continue.dev connects to Ollama at http://localhost:11434 by default. Diagnose connection issues:

# Verify Ollama is running
curl http://localhost:11434
# Expected: "Ollama is running"

# List pulled models
ollama list

# Pull a model if not present
ollama pull llama3.2
ollama pull starcoder2:3b  # for tab autocomplete

config.json for Ollama:

{
  "models": [
    {
      "title": "Llama 3.2",
      "provider": "ollama",
      "model": "llama3.2"
    }
  ]
}

If Ollama runs in Docker: set OLLAMA_HOST=0.0.0.0 and update apiBase to http://host.docker.internal:11434.

Tab autocomplete not triggering

Tab autocomplete needs a separate fast model — using a large chat model for autocomplete causes slow, poor-quality suggestions. Configure a dedicated autocomplete model:

{
  "tabAutocompleteModel": {
    "title": "Starcoder2 3B",
    "provider": "ollama",
    "model": "starcoder2:3b"
  },
  "tabAutocompleteOptions": {
    "debounceDelay": 500,
    "maxPromptTokens": 1500,
    "multilineCompletions": "auto"
  }
}

Also check VS Code settings: search "Continue" → ensure Enable Tab Autocomplete is checked. Autocomplete triggers after a short pause when typing in code files — it won't trigger in comments or plain text files.

API key errors (Anthropic / OpenAI / OpenRouter)

Each provider uses different field names in config.json. Ensure you're using the correct format:

{
  "models": [
    {
      "provider": "anthropic",
      "model": "claude-3-5-sonnet-20241022",
      "apiKey": "sk-ant-..."     // Anthropic key
    },
    {
      "provider": "openai",
      "model": "gpt-4o",
      "apiKey": "sk-..."          // OpenAI key
    },
    {
      "provider": "openrouter",
      "model": "anthropic/claude-3.5-sonnet",
      "apiKey": "sk-or-v1-...",   // OpenRouter key
      "apiBase": "https://openrouter.ai/api/v1"
    }
  ]
}

Never commit config.json with API keys to version control — add ~/.continue/config.json to .gitignore.

JetBrains plugin issues

Continue supports IntelliJ, PyCharm, WebStorm, and other JetBrains IDEs via a separate plugin:

1. 1
   Install from JetBrains Marketplace: Search "Continue" in the JetBrains Marketplace (Settings → Plugins → Marketplace) — do not use the VS Code extension.
2. 2
   Same config.json: The JetBrains plugin reads the same ~/.continue/config.json as the VS Code extension — no separate configuration needed.
3. 3
   Missing tool window: If the Continue panel is missing: View → Tool Windows → Continue. If not listed, the plugin may not have activated yet.
4. 4
   Restart the IDE: Restart the IDE after installing — the Continue plugin does not activate until a full IDE restart.
5. 5
   Plugin version lag: The JetBrains plugin may lag behind the VS Code version. Check the plugin changelog for known issues and minimum IDE version requirements.