Configuration

Configuring Sleev means running the local gateway and telling your harnesses which provider traffic should use. The setup wizard in the sleev TUI handles the common path. Manual harness config examples and setup instructions are outlined in Harness Setup.

Prerequisites

Before you configure a harness, make sure the gateway is installed and running. The Quickstart walks through this in full. You should have:

  1. Installed the sleev command-line tool.
  2. Run setup to register this machine, install the gateway, and start it.
  3. Confirmed the gateway is running locally on 127.0.0.1:17321.

You can check that the gateway is up at any time with:

Setup wizard

The easiest way to configure any harness is the setup wizard built into the TUI, and it works the same for every supported harness.

  1. In the TUI, open the Config Wizard. It lists the harnesses detected on your machine.
  1. Select the harness you want to configure. sleev shows the providers that we support for that harness.
  2. Enable the provider you want to use. sleev writes the change into your harness's config for you. The manual configuration option in the TUI links to Harness Setup if you prefer to edit the harness config file yourself.

Command-line setup

To configure a harness from the CLI, open the harness picker directly:

This opens the same harness and provider picker without launching the full TUI.

For the full command list, see the CLI Reference.

Providers

Built-in provider support depends on the harness. The table below lists the providers Sleev supports for each harness. If you need a compatible provider that is not listed, use a custom base URL instead.

Claude Code sleeve-harness: claude code
anthropic
Anthropic Claude models, with a pay-as-you-go API key or a Claude Max subscription.
Codex CLI sleeve-harness: codex cli
openai
OpenAI models with a pay-as-you-go API key.
codex
OpenAI's Codex backend through ChatGPT sign-in, with no separate API key.
OpenCode sleeve-harness: opencode
anthropic
Anthropic Claude models with a pay-as-you-go API key. Claude Max subscriptions are not supported here.
openai
OpenAI models with a pay-as-you-go API key.
codex
OpenAI's Codex backend through ChatGPT sign-in, with no separate API key.
copilot
GitHub Copilot models through the Copilot API.
google
Google Gemini models through the Gemini API.
moonshotai
Moonshot AI's general API for Kimi models.
kimi-coding
Moonshot's Kimi coding plan subscription.
mimo
Xiaomi MiMo Token Plan Singapore using the OpenAI-compatible /v1 API.
zai-coding-plan
Z.AI's coding plan API for GLM coding models.
opencode
OpenCode Zen.
opencode-go
OpenCode Go.

Custom providers

Use custom providers when your endpoint uses an OpenAI-compatible, Anthropic-compatible, or Gemini-compatible format but is not one of Sleev's built-in providers. Most users should leave this unset.

Keep your harness pointed at the local Sleev gateway, then send sleeve-harness and sleeve-base-url in the harness's request headers. The exact syntax depends on the harness; see the Claude Code, Codex CLI, and OpenCode examples in Harness Setup for how each config sets request headers. Do not set sleeve-provider for custom providers:

Custom headers 
sleeve-harness: opencode
sleeve-base-url: https://provider.example.test/v1

Use a full http or https base API URL. Include a path such as /v1 only when that is part of the provider's API base.

Gateway bind address

By default, the Sleev gateway listens on 127.0.0.1:17321, so only harnesses on the same machine can reach it.

Use sleev gateway bind when a harness runs on another trusted machine and needs to connect to this gateway over your local network:

Use --port if you want a different gateway port:

To return to the default local-only address, reset the bind setting:

If the harness is on another machine, use the gateway bind address as the harness base URL, for example http://<gateway-host>:17321.

Only bind to 0.0.0.0 on networks you trust. Other machines that can reach your computer may be able to reach the gateway too.

Limited internet

Most users sign in through the TUI before configuring their harness. If the target machine cannot use normal browser sign-in, or if Sleev should avoid sharing tracking data and usage patterns, accounts approved for offline licensing can activate the local gateway with an offline license instead.

Note: offline licensing is currently only supported for enterprise customers. To request access, contact support.

Once the gateway is licensed, the harness configuration above is the same: your harness still points at the local Sleev gateway and sends sleeve-harness plus either sleeve-provider or sleeve-base-url with each request.

Deconfigure

To remove Sleev from this machine, sign out:

Signing out removes local auth state and gateway config, deconfigures supported harnesses, and stops the local gateway service.

You can also do this from the sleev TUI by signing out. If you only want to stop routing one harness through Sleev, remove the Sleev-specific base URL and headers shown in Harness Setup from that harness's config file.