Contents
To use SolCard Inference in Cline, choose “OpenAI Compatible” as the API Provider, set the Base URL to https://api.solcardinference.ai/v1, paste your sk-si- key, and enter solcard-inference-fast as the model. Then fill in Model Configuration: max output 32,768 tokens, image support off. The field names below come from Cline’s own docs; we haven’t run Cline ourselves, but the API behind them is the same one OpenCode and Aider use.
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.
- Cline in your editor.
The provider settings
Open Cline’s settings and fill in:
| Field | Value |
|---|---|
| API Provider | OpenAI Compatible |
| Base URL | https://api.solcardinference.ai/v1 |
| API Key | your sk-si-... key |
| Model | solcard-inference-fast |
The Base URL is the step that matters. Cline’s docs call it crucial: it must be your provider’s endpoint, not https://api.openai.com/v1. Ours ends in /v1, with nothing after it.
Model Configuration
Cline also asks for the model’s limits and prices, so it can manage context and estimate cost. For our models:
| Setting | Fast | Large | Base |
|---|---|---|---|
| Max Output Tokens | 32,768 | 32,768 | 32,768 |
| Context Window | 256,000 | 256,000 | 256,000 |
| Image Support | Off | Off | Off |
| Input Price (per 1M) | $1.25 | $3.75 | $1.25 |
| Output Price (per 1M) | $2.50 | $6.25 | $3.75 |
Limits from the models page, prices from the pricing page.
- Max Output Tokens: every model is capped at 32,768 per request. Don’t set it higher.
- Image Support: leave it off. Our models take text only, and an image in a request is rejected.
- Computer Use: Cline describes this as for models with tool or function calling. All three models call function tools, but we haven’t tested this setting in Cline.
- Prices: cached input bills at a tenth of the input rate ($0.125 on Fast). Your usage page has the real cost of every request, whatever Cline estimates.
Which model to pick
solcard-inference-fastfor most tasks: reading code, small edits, running tests.solcard-inference-largefor planning a change or a bug Fast can’t find. It always reasons and costs about three times as much.solcard-inference-basefor a second opinion.
To switch, change the Model field and its Model Configuration values together. Fast, Large or Base has the details.
Check it works
Before you open Cline, check the key and URL from a terminal:
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. If that works, give Cline a small task, such as “Explain what this file does.” A reply means the settings are right; an error means one of the fields below is off.
What a task costs
A Cline task is many requests, and each one resends the conversation so far. Most of that resent input is cached, at a tenth of the normal rate.
Say one request on Fast resends 60,000 tokens, 55,000 of them cached, and writes 1,200 tokens:
- 55,000 cached × $0.125 per million = $0.0069
- 5,000 new × $1.25 per million = $0.0063
- 1,200 output × $2.50 per million = $0.0030
- About $0.016 a request
A task of 20 such requests is about $0.32 on Fast. On Large, the same request is about $0.047, plus its reasoning, which bills as output.
If you use Cline every day, look at the plans. Starter is $20 a month for 20% more usage than pay as you go, and raises the limit from 60 to 120 requests a minute.
Troubleshooting
These are the API’s responses; Cline may word them differently.
- 404, “Cannot POST /chat/completions”: the Base URL is missing
/v1. - 401
invalid_api_key: the key is wrong, revoked or has a stray space. Cline’s docs also list “Invalid API Key” for this case. - 400
model_not_found: the Model field doesn’t match an id. Cline’s docs list this as “Model Not Found”. - 400
unsupported_content: an image reached the model. Turn Image Support off and don’t attach images.
The errors page lists every code.
FAQ
Can I use a custom base URL in Cline? Yes. That’s what the OpenAI Compatible provider is for: any API that speaks OpenAI’s format, at the Base URL you give it.
Why set Max Output Tokens to 32,768? That’s the most any of our models writes in one request.
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 with Claude Code or Codex too? Yes, with one sentence: see Claude Code and Codex.