RealRelay.ai
Coding agents

Cherry Studio

Connect Cherry Studio to RealRelay.ai as an OpenAI-type provider: import the key from the console or add it by hand with the site root as the API address, then check the connection and add models.

Create an API keyBrowse the catalog
Protocol
OpenAI Chat Completions
Base URL
https://www.realrelay.ai
Official docs
docs.cherry-ai.com
Checked against the tool’s docs
2026-10
Experimental. Not yet run end to end against RealRelay.ai. The steps follow the tool’s own documentation; if one fails, contact support with the error.

Overview

Cherry Studio is a desktop chat client that works with many model providers at once, with assistants, knowledge bases and tools on top. It connects to RealRelay.ai as an OpenAI-compatible provider, which uses the /v1/chat/completions endpoint.

  • An API key from the API keys page of the console.

  • The API address is the site address without /v1. Unlike most OpenAI-compatible clients, Cherry Studio appends /v1/chat/completions itself.

  • A model ID exactly as GET /v1/models lists it, for example z-ai/glm-5.2.

Cherry Studio's own docs still use the 1.x names for some buttons, and the 2.x app renames a few of them. This page gives both where they differ. The setup has not yet been run end to end against RealRelay.ai.

One-click setup

The console can send a key straight to Cherry Studio through a cherrystudio:// link. Install Cherry Studio first (see Install below) so your browser knows how to open it.

  1. 1

    Open the key's menu

    On the API Keys page, open the ⋯ menu in the Actions column of the key you want to use, then choose Chat → Cherry Studio. Allow the browser to open the app.

  2. 2

    Confirm the key

    Cherry Studio opens its provider settings and asks whether to add the API key for its built-in New API provider. The dialog shows the provider ID and the base URL, which is the site root. Click Add.

  3. 3

    Add models and turn the provider on

    The link brings the address and key, not the models. Add models and switch the provider on as described under Configure and Models below.

If your console has no Chat → Cherry Studio item, the manual steps below work just as well.

Install

Download Cherry Studio from the official download page or from GitHub Releases. It runs on Windows, macOS and Linux.

Pick the build for your Mac's chip: Apple Silicon for M-series Macs, Intel otherwise. You can check in the Apple menu under About This Mac. Open the downloaded file and drag Cherry Studio into Applications.

Run the installer. Windows 7 is not supported. If the installer asks for the Microsoft Visual C++ Redistributable, let it install it.

Download the Linux package for your architecture from GitHub Releases.

The install worked if Cherry Studio opens to its main window.

Configure

  1. 1

    Add a provider

    Open Settings (the gear icon) → Model Provider (Model Services in 1.x). Click Add below the provider list, name the provider, set the type to OpenAI, and save.

    Settings → Model Provider → Add
    Provider Name: RealRelay.aiProvider Type: OpenAI
  2. 2

    Enter the key and address

    Open the new provider. Paste your key into API Key, and put the site root into the address field, called API Host in 2.x and API Address in 1.x. Replace sk-*** with your key.

    Settings → Model Provider → RealRelay.ai
    API Key:  sk-***API Host: https://www.realrelay.aiModel ID: z-ai/glm-5.2
  3. 3

    Turn the provider on

    Switch on the toggle at the top right of the provider page. A provider that is off keeps its settings, but its models do not appear in the model picker.

  • Do not add /v1 to the address. Cherry Studio adds /v1/chat/completions, so /v1 in the field becomes /v1/v1/…. The address preview under the field shows the final URL.

  • An address ending in # is used exactly as typed, with nothing appended. You only need this if you want to spell out the full endpoint:

    API Host (used exactly as typed)
    https://www.realrelay.ai/v1/chat/completions#

  • The API Key field takes several keys separated by English commas. Cherry Studio uses them in turn.

  • If you used the console's one-click setup, the settings are on the built-in New API provider instead; you do not need a second provider.

Verify

  1. 1

    Call the endpoint directly

    Run this with your key. A JSON reply with a choices array means the address, key and model work.

    Terminal
    1curl -sS "https://www.realrelay.ai/v1/chat/completions" \2  -H "Authorization: Bearer sk-***" \3  -H "content-type: application/json" \4  -d '{"model": "z-ai/glm-5.2", "max_tokens": 16, "messages": [{"role": "user", "content": "Reply with OK"}]}'
  2. 2

    Run Cherry Studio's check

    Add at least one model (see Models), then click Check next to the API Key field. Cherry Studio sends a test request with one of the provider's models and reports whether it succeeded.

  3. 3

    Send a message

    Go back to the chat page, pick the RealRelay.ai model in the model selector, and send a short message. The request then appears in the console's usage logs.

Models

Click the model list button at the bottom of the provider page, Sync models in 2.x or Manage in 1.x. Cherry Studio loads the list from RealRelay.ai. Click + next to each model you want; only models you add show up in the chat model picker. You can also add a model by typing its ID.

Terminal
curl -sS "https://www.realrelay.ai/v1/models" \  -H "Authorization: Bearer sk-***"

  • Check the model ID shown under each name before you add it. RealRelay.ai lists IDs with a channel prefix, such as z-ai/glm-5.2, and similar names from different channels are different models.

  • Cherry Studio does not need a context window or output limit for chat. Assistants have their own settings for how many earlier messages to send.

  • For tools and MCP servers, choose a model with tool calling. Most coding models in the catalog have it.

For help picking a model, see Choosing a model.

Troubleshooting

  • 404. The address ends in /v1 or /v1/chat/completions. Keep only the site root.

  • 401 or an invalid key message. The key is wrong, disabled or out of quota. Copy it again from the API keys page, without spaces.

  • The model does not appear in the chat picker. The provider is switched off, or the model was listed but not added. Turn the provider on and click + next to the model.

  • Check fails although chat works. The check uses one of the provider's models. Remove models that your key cannot call, or that are not chat models.

  • A "model not found" error. The model ID does not match RealRelay.ai's list exactly, often a missing prefix. Delete the model and add it again from the synced list.

  • Nothing happens after clicking Cherry Studio in the console. Cherry Studio is not installed or the browser blocked the link. Start Cherry Studio once, then click again, or use the manual steps.