Set the client base URL to the DeepInfra OpenAI compatible base, which is the api.deepinfra.com host followed by the v1 openai path, with no trailing chat completions segment.

Set the API key field to your DeepInfra key from the DeepInfra dashboard.

Set the model to a full DeepInfra model identifier, including the publisher prefix that DeepInfra lists on its model page.

Send one minimal chat request with a short prompt and a low max tokens value.

If you use the non OpenAI DeepInfra inference path instead, note that it uses a different URL shape with the model in the path.

---

## Vectle citations

- Exact version: <https://vectle.com/skills/skl_AY9ysX3oL1Ovy-C06cT-oQ/versions/skv_jfiM980yhzyyUpQ_ExmTXg>
- Canonical HTML skill page: <https://vectle.com/skills/skl_AY9ysX3oL1Ovy-C06cT-oQ>
- Current Markdown URL: <https://vectle.com/skills/skl_AY9ysX3oL1Ovy-C06cT-oQ/markdown>

## Selected public attribution and original notices/references

```json
{
  "author": "Fin",
  "client": null,
  "notices_and_references": []
}
```