Ir al contenido
Nexus Docs
↑↓Navigate↵SelectescClose

Cline (VS Code) → Nexus

Route the Cline VS Code extension through the Nexus OpenAI-compatible edge

Last updated on:

Esta página aún no está disponible en tu idioma.

Route the Cline VS Code extension through the Nexus OpenAI-compatible edge (/v1). Cline’s OpenAI Compatible provider calls chat completions; Nexus verifies a gateway API key (nxs_…) and dispatches to the configured upstream.

Shared prerequisites (gateway URL, nxs_… key, registry model, BYOK/PTB): Client Integrations.

1. Open Cline settings

Section titled “1. Open Cline settings”
  1. Install the Cline extension in VS Code / Cursor.
  2. Open the Cline sidebar → gear (settings).
  3. Set API Provider to OpenAI Compatible.

2. Fill the provider fields

Section titled “2. Fill the provider fields”
Field Value
Base URL https://api.example.com/v1 (local: http://127.0.0.1:14450/v1)
API Key nxs_… gateway secret (no Bearer prefix)
Model ID Exact id from GET /v1/models (for example openai/gpt-5.4-mini)

Do not append /chat/completions to the Base URL. Cline adds that path.

Model configuration

Section titled “Model configuration”

Cline cannot infer limits for an arbitrary compatible endpoint. Set these honestly from the model’s registry / provider docs:

Setting Guidance
Context window Match the model’s context size
Max output tokens A safe completion cap the model accepts
Image support On only if the model advertises vision
Computer use / tools On when you want agent tool calls (gateway forwards OpenAI tool fields)

Optional input/output price fields only affect Cline’s own cost display; Nexus billing still uses the registry rate card.

3. Verify

Section titled “3. Verify”

Use Cline’s Verify control when present, or send a short Plan/Act task.

Independent check:

Terminal window
GATEWAY_URL="${GATEWAY_URL:-https://api.example.com}"
NEXUS_API_KEY='nxs_…'


curl -sS "$GATEWAY_URL/v1/models" \
  -H "Authorization: Bearer $NEXUS_API_KEY" | jq .


curl -sS "$GATEWAY_URL/v1/chat/completions" \
  -H "Authorization: Bearer $NEXUS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-5.4-mini",
    "messages": [{"role": "user", "content": "ping"}],
    "max_tokens": 16
  }' | jq .

Common failures

Section titled “Common failures”
Symptom Likely cause
Invalid API key Typo, whitespace, or nxd_… / upstream sk-… pasted instead of nxs_…
Model not found Model id not in the tenant catalog, or Base URL missing /v1
Connection errors Gateway hostname unreachable from the IDE host
Tools / agent steps fail Model Configuration flags off, or selected model lacks tool support upstream

Cline’s OpenAI Compatible path targets chat completions. If you need the Responses API agent loop, use Codex or another client that speaks /v1/responses — Nexus exposes that route on the same gateway.

Section titled “Related”