Use SolCard Inference in Continue

To use SolCard Inference in Continue, add a model to ~/.continue/config.yaml with provider: openai, apiBase: https://api.solcardinference.ai/v1, your sk-si- key and solcard-inference-fast as the model. Continue’s OpenAI provider docs show exactly this pattern for any OpenAI-compatible API. We haven’t run Continue ourselves; every field below comes from Continue’s docs and ours.

What you need

• A SolCard Inference API key. Make one on the dashboard’s API keys page. Keys start with sk-si- and are shown once.
• Some balance or a plan. One prepaid USD balance pays for every key and model; see how billing works.
• Continue in VS Code or a JetBrains IDE.

Put the key in a .env file

Continue’s IDE extensions can’t read your shell’s environment variables. Its FAQ says so directly, so export in a terminal won’t help. Instead, Continue reads ${{ secrets.NAME }} references from .env files, in this order: the workspace .env, the workspace .continue/.env, then the global ~/.continue/.env.

For a key you use in every project, put it in the global file:

# ~/.continue/.env
SOLCARD_INFERENCE_API_KEY=sk-si-...

Continue’s FAQ says to restart the IDE after you add or change a secret. If you use a workspace .env, keep it out of git.

The config.yaml block

The local config lives at ~/.continue/config.yaml on macOS and Linux, and %USERPROFILE%\.continue\config.yaml on Windows. Per Continue’s configuration docs, you can also open it from the configs dropdown at the top right of the chat input: click the cog next to “Local Config”.

Add these entries under models:. If the file is new, keep the name, version and schema lines at the top:

name: My Config
version: 0.0.1
schema: v1

models:
  - name: SolCard Inference Fast
    provider: openai
    model: solcard-inference-fast
    apiBase: https://api.solcardinference.ai/v1
    apiKey: ${{ secrets.SOLCARD_INFERENCE_API_KEY }}
    capabilities:
      - tool_use
    defaultCompletionOptions:
      contextLength: 1048576
      maxTokens: 32768

  - name: SolCard Inference Large
    provider: openai
    model: solcard-inference-large
    apiBase: https://api.solcardinference.ai/v1
    apiKey: ${{ secrets.SOLCARD_INFERENCE_API_KEY }}
    roles:
      - chat
    capabilities:
      - tool_use
    defaultCompletionOptions:
      contextLength: 1000000
      maxTokens: 32768

  - name: SolCard Inference Base
    provider: openai
    model: solcard-inference-base
    apiBase: https://api.solcardinference.ai/v1
    apiKey: ${{ secrets.SOLCARD_INFERENCE_API_KEY }}
    roles:
      - chat
    defaultCompletionOptions:
      contextLength: 262144
      maxTokens: 32768

What each field does, per Continue’s config reference:

Every model writes at most 32,768 tokens per request, so keep maxTokens at that or lower. Context is the combined input and output budget, per our models page.

Leave out image_input. Our models take text only, and a request with an image is rejected with unsupported_content.

Continue’s docs say it uses OpenAI’s /responses endpoint by default only for o-series and gpt-5 models. Our model ids are neither, so the block above should go to /chat/completions, which our API serves.

Which model to pick

• solcard-inference-fast for chat, inline edits and applying changes. It’s the cheapest, so it gets the default roles above.
• solcard-inference-large for planning a change or a bug Fast can’t find. It always reasons and costs about three times as much, so the block keeps it to chat.
• solcard-inference-base for a second opinion in chat.

Don’t use any of them for autocomplete. Continue’s autocomplete docs say chat models “will often perform poorly even with extensive prompting” at it, and recommend small models trained for that job. That’s why none of the entries above lists autocomplete as a role. Fast, Large or Base has more on the three models.

Check it works

Check the key and URL from a terminal first:

curl https://api.solcardinference.ai/v1/models \
  -H "Authorization: Bearer sk-si-..."

You should get a list with solcard-inference-fast, solcard-inference-large and solcard-inference-base. Then pick “SolCard Inference Fast” in Continue’s model dropdown and ask something small, such as “Explain what this file does.” A reply means the config is right.

What a chat turn costs

Continue sends your question plus the code you give it as context, so input is most of each request. Prices are USD per 1M tokens, from the pricing page:

Say a turn on Fast sends 20,000 tokens, 15,000 of which bill as cached input, and writes 1,000 tokens:

• 15,000 cached × $0.125 per million = $0.0019
• 5,000 new × $1.25 per million = $0.0063
• 1,000 output × $2.50 per million = $0.0025
• About $0.011 a turn

The same turn on Large is about $0.031, plus its reasoning, which bills as output.

If you use Continue all day, look at the plans. Starter is $20 a month for 20% more usage than pay as you go, with a limit of 120 requests a minute. Plan usage resets every 5 hours and every week, and doesn’t roll over.

Troubleshooting

These are the API’s error codes, from our errors page; Continue may word them differently.

• 401 invalid_api_key: the key is missing, revoked or unknown. Check the .env file name and the secret name, then restart the IDE.
• 400 model_not_found: the model field doesn’t match an id. Use solcard-inference-fast, -large or -base.
• 400 unsupported_content: an image or file reached the model. Our models take text only.
• 402 insufficient_balance: credits and plan can’t cover the request. Top up and retry.
• Agent mode is grayed out: Continue’s FAQ says Agent mode needs tool_use. Check that the model you picked has capabilities: [tool_use].

FAQ

Can I use a custom OpenAI-compatible API in Continue? Yes. Set provider: openai and point apiBase at the API. Continue’s docs show it for any OpenAI-compatible provider.

Why doesn’t my key load from my shell? Continue’s VS Code and JetBrains extensions can’t read shell environment variables. Put the key in a .env file Continue reads, such as ~/.continue/.env.

Can I keep my other models in Continue? Yes. These entries sit next to the others under models:, and you switch in the model dropdown.

Are my prompts or code stored? No. Prompts and responses aren’t stored or logged, on any model. Keeps nothing follows one request through.

Does it work in Cline or Claude Code too? Yes. See Set up SolCard Inference in Cline, or Claude Code and Codex for the one-sentence setup. For VS Code’s built-in chat, see Use SolCard Inference in VS Code with Custom Endpoint.