nodecode-opper

Opper: an EU-hosted AI gateway, 700+ models from 30+ providers behind one OpenAI-compatible API. A Nodecode rewrite of opper-ai/hermes-opper by opper-ai.

With it, /connect offers Opper, /models lists what the gateway publishes, and a turn on opper/<model> goes to https://api.opper.ai/v3/compat over the chat-completions wire, the way the plugin's profile sent it.

Get it

Tell the agent "add the opper cell", or run nodecode add nc://opper.

Set it up

Make a key at https://platform.opper.ai, then save it with /connect or set OPPER_API_KEY. A saved key wins over the variable.

Optional, in ~/.nodecode/config.jsonc:

{
  "opper": {
    // where the gateway is served; the chat lane appends /chat/completions
    "base_url": "https://api.opper.ai/v3/compat"
  }
}

OPPER_BASE_URL in the environment stands in for that setting, the way the plugin's own env_vars allowed.

Use it

Model ids are the gateway's pool names: claude-sonnet-4-6, claude-opus-5, gpt-5.5, gpt-5.4-mini, gemini-3.8-flash, deepseek-v4-pro, kimi-k3, mistral-large-2512. A pool is every provider serving that model, and the gateway picks the route per request. Prefix a vendor to pin one: anthropic/claude-sonnet-4-6, azure/gpt-5.5. The catalogue at https://api.opper.ai/v3/compat/models is what /models reads, so the ids on offer are the live ones; the eight above are the roster shown when that address cannot be reached.

Switch mid-session with /model opper:gpt-5.4-mini.

Test

From this repository's root:

.github/test-cell.sh providers/nodecode-opper

MIT licensed. The original plugin and its licence are named in NOTICE.