Integrations

1. 01Open Settings (⌘⇧J on macOS, Ctrl+Shift+J on Windows/Linux) and go to Models.
2. 02Paste your Selora key into OpenAI API Key and enable Override OpenAI Base URL.
3. 03Add each model id under Models → Add model, then select it in chat.

OpenAI API Key:             sk-gw-...
Override OpenAI Base URL:   https://api.selora.lol/v1
Add model:                  deepseek-v4.1-flash-1017
                            kimi-k3.1-pro
                            claude-sonnet-4-5

Cursor's own hosted models stay on Cursor's billing. Only requests sent through the override are billed to your Selora key. If your build has no override field, install Cline inside Cursor and point that at the same endpoint.

1. 01Install Cline from the VS Code Marketplace (or your fork’s extension panel).
2. 02Open the Cline panel → gear icon → API Provider: OpenAI Compatible.
3. 03Fill in the three fields below and click Verify.

API Provider:  OpenAI Compatible
Base URL:      https://api.selora.lol/v1
API Key:       sk-gw-...
Model ID:      claude-sonnet-4-5

“Invalid API Key” means the key was mistyped; “Model Not Found” means the Model ID is not one the gateway serves. Run the model check below to list valid ids.

1. 01Install the Roo Code extension, then open its settings (gear icon).
2. 02Set API Provider to OpenAI Compatible.
3. 03Enter the base URL and key, then pick or type a model id.

API Provider:  OpenAI Compatible
Base URL:      https://api.selora.lol/v1
API Key:       sk-gw-...
Model ID:      kimi-k3.1-pro

Roo Code keeps its own context-window and pricing metadata per model. Set Context Window / Max Output Tokens to match the model you chose — the values are on the Models page.

1. 01Install OpenCode: curl -fsSL https://opencode.ai/install | bash
2. 02Add the provider block to ~/.config/opencode/opencode.json (or a project opencode.json).
3. 03Export SELORA_API_KEY, launch opencode, then pick selora/<model> with /models.

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "selora": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Selora",
      "options": {
        "baseURL": "https://api.selora.lol/v1",
        "apiKey": "{env:SELORA_API_KEY}"
      },
      "models": {
        "deepseek-v4.1-flash-1017": { "name": "DeepSeek V4.1 Flash" },
        "kimi-k3.1-pro": { "name": "Kimi K3.1 Pro" },
        "claude-sonnet-4-5": { "name": "Claude Sonnet 4.5" }
      }
    }
  },
  "model": "selora/kimi-k3.1-pro"
}

export SELORA_API_KEY="sk-gw-..."
cd /path/to/project
opencode

Keep the key out of the file: set SELORA_API_KEY in your shell and reference it with {env:SELORA_API_KEY}.

1. 01Open ~/.continue/config.yaml (Continue 1.x; the legacy config.json uses the same fields in JSON).
2. 02Add one entry per model you want, with apiBase pointing at the gateway.
3. 03Reload the Continue panel — the models appear in the picker.

name: selora
version: 1.0.0
schema: v1
models:
  - name: Selora · Claude Sonnet 4.5
    provider: openai
    model: claude-sonnet-4-5
    apiBase: https://api.selora.lol/v1
    apiKey: sk-gw-...
    roles: [chat, edit, apply]
  - name: Selora · DeepSeek V4.1 Flash
    provider: openai
    model: deepseek-v4.1-flash-1017
    apiBase: https://api.selora.lol/v1
    apiKey: sk-gw-...
    roles: [autocomplete]

Do not append /chat/completions to apiBase — Continue adds the path itself. A 404 here almost always means the base URL ends in the wrong place.

1. 01Install: npm install -g @openai/codex
2. 02Add the provider block to ~/.codex/config.toml.
3. 03Export SELORA_API_KEY and launch with codex --profile selora.

[model_providers.selora]
name = "Selora"
base_url = "https://api.selora.lol/v1"
env_key = "SELORA_API_KEY"
wire_api = "responses"
requires_openai_auth = false

[profiles.selora]
model = "deepseek-v4.1-flash-1017"
model_provider = "selora"

export SELORA_API_KEY="sk-gw-..."
cd /path/to/project
codex --profile selora

wire_api = "responses" is what current Codex builds require for custom providers and targets /v1/responses. "chat" still works only on older pinned builds — treat it as legacy. Reserved provider names (openai, ollama, lmstudio) take precedence and cannot be overridden.

1. 01Add the endpoints.custom block to librechat.yaml.
2. 02Set SELORA_API_KEY in the same .env LibreChat loads.
3. 03Start with CONFIG_PATH=./librechat.yaml so the file is read.

version: 1.2.8
cache: true
endpoints:
  custom:
    - name: "Selora"
      apiKey: "${SELORA_API_KEY}"
      baseURL: "https://api.selora.lol/v1"
      models:
        default:
          - "claude-sonnet-4-5"
          - "kimi-k3.1-pro"
          - "deepseek-v4.1-flash-1017"
        fetch: false
      titleConvo: true
      titleModel: "deepseek-v4.1-flash-1017"
      modelDisplayLabel: "Selora"

fetch: false keeps the dropdown to the three ids above instead of probing the gateway on boot. Drop the field if you would rather list models live.

1. 01Open AI Response Configuration → Chat Completion.
2. 02Set Chat Completion Source to Custom (OpenAI-compatible).
3. 03Paste the endpoint and key, click Connect, then pick a model.

Chat Completion Source:  Custom (OpenAI-compatible)
Custom Endpoint:         https://api.selora.lol/v1
Custom API Key:          sk-gw-...
Model:                   claude-sonnet-4-5

The endpoint is the /v1 root — SillyTavern appends /chat/completions itself. If the model dropdown stays empty after Connect, refresh it and type the id by hand.