# AI Provider Setup and Notifications ### **Supported AI Providers** Nimbalyst is meant to work with your coding agents. The following are currently supported: * Claude Code * Codex (SDK) * Codex over ACP (Alpha) * OpenCode (Alpha) * GitHub Copilot (Alpha) Alpha providers are toggled on individually from their own settings panels. See [Alpha Features](/setup-nimbalyst/alpha-features.md). You can configure Nimbalyst to also access AI Models directly. ### OpenAI Codex You must separately download the Codex CLI (or app). Either authenticate there in the CodexCLI or provide Nimbalyst with an API token for OpenAI.. On the Global Settings screen (Select Nimbalyst -> Settings), enable Codex and configure its integration with Nimbalyst. Codex transcripts have feature parity with Claude in Nimbalyst: * **Inline edit cards** — Codex `file_change` tool calls render as red/green edit cards in the transcript, matching how Claude's Edit tool already renders. * **Slash command autocomplete** — Codex slash commands surface in the same unified `/` picker as Claude Code commands. * **Cross-agent skills and commands** — Skills and commands written for one agent run in the other, so workflow discovery is unified across providers. * **Reasoning blocks** — Codex reasoning items map into transcript thinking blocks. ### Claude Code
On the Global Settings screen (Select Nimbalyst -> Settings), enable Claude Code and configure its integration with Nimbalyst. New users default to **Opus 4.6 with 1M context window**. Existing Opus users are automatically migrated to this configuration. If you already have a coding agent installed and running on your desktop, we will use it through its SDK. If you do not have Claude Code CLI, we will install it for you automatically. You must either enter your API Key in the Settings in Nimbalyst or authenticate to Anthropic in the CLI. **Claude Enterprise SSO** is supported — if your organization uses Claude Enterprise, you can authenticate via SSO. #### Effort Level Claude Code supports an effort level slider that maps to the Claude Code CLI's `/effort` setting. Higher levels let the model spend more thinking budget on a turn at the cost of more tokens. Pick from five levels: * **low** * **medium** * **high** * **xhigh** * **max** The Effort Level selector is in the agent's session controls. The choice is forwarded to Claude Code via the `CLAUDE_CODE_EFFORT_LEVEL` environment variable. ### OpenAI Codex over ACP (Alpha) An alternative way to run Codex inside Nimbalyst. Instead of using the Codex SDK, this provider speaks the Agent Client Protocol over stdio, which gives Nimbalyst native pre/post file-edit hooks. The practical benefit is more accurate diff baselines (especially for newly created files) and deterministic per-session attribution for `apply_patch` edits. Your MCP servers are forwarded to the agent automatically. Enable it under **Settings → Agent Providers** by selecting the **Codex (ACP)** provider. It appears in the model picker alongside the SDK Codex provider. ### OpenCode (Alpha) OpenCode is an open-source coding agent with multi-model support — it works with Claude, OpenAI, Gemini, and local models through a unified interface. Select **OpenCode** under **Settings → Agent Providers**: 1. Install the OpenCode CLI (use the **Install OpenCode CLI** button or follow the OpenCode docs). 2. Toggle **Enable OpenCode** on. 3. Optionally provide an API key — model selection and provider configuration are managed through OpenCode's own settings, so no additional configuration is required here.

OpenCode agent provider settings

### GitHub Copilot (Alpha) GitHub Copilot is supported as a coding agent via the ACP (Agent Communication Protocol) server mode. It uses your existing Copilot CLI login for authentication. Select **GitHub Copilot** under **Settings → Agent Providers**: 1. Install the Copilot CLI (`npm install -g @github/copilot`, or use the **Install Copilot CLI** button). 2. Toggle **Enable GitHub Copilot** on. 3. Authenticate by running `copilot` and using the `/login` command. Model selection is managed by Copilot — no additional API key is required.

GitHub Copilot agent provider settings

### Integrate other AI Models with Nimbalyst (optional) * Nimbalyst can integrate with Anthropic, OpenAI or LM Studio for Local Models * You must provide your API key to integrate with each * As these are not coding agents, only some features will be supported * For LM Studio, you must have this installed locally * (Return to this screen from Nimbalyst -> Settings if you want to change this config later) ### Notification that AI response is complete Under Nimbalyst -> Settings. Click on Notifications
By default: * You are notified when you AI session completes * And if you are not in the window, it will pop up a notification when the session completes. Clicking on that notification takes you to the window where the session was running. You can turn one or both off --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.nimbalyst.com/setup-nimbalyst/ai-provider-setup-and-notifications.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.