無制限 Claude で Cherry Studio を動かす

Cherry Studio は長く探索的なチャットを誘います——会話を開いたまま、問い続ける。このスタイルは着実にトークンを消費し、まさにそこで定額・無制限の Claude プランが効いてきます。好きなだけチャットしても、料金は横ばいのままです。

Add the provider once

Open Settings, then Model Services, and add a provider. With an unlimited plan this is a one-time setup that frees you from ever thinking about per-message cost in Cherry Studio again.

Pick the type when prompted. For Claude on the unlimited gateway you want the Anthropic type, which speaks the native Messages format the gateway serves.

Because Cherry Studio is a desktop client meant for daily, sustained use, pairing it with a flat-rate plan fits how people actually use it — open all day, dipped into constantly. One provider entry set once is all it takes to make that whole pattern a fixed cost. An all-day client and a flat rate are a natural pairing, and the one-time setup is what seals it. After that first configuration you simply chat, and the plan absorbs whatever volume the day's conversations bring.

Anthropic type, your key

Choose Anthropic, enable it, and paste your unlimited key into the API Key field. That single key backs every conversation you start, no matter how many messages each one runs to.

Because the provider is native Anthropic, there is no protocol toggle to manage. Set the type and the client sends Claude requests correctly from then on.

There is nothing to rotate per session or per project — the same unlimited key stays in place for the life of the plan. Once it is entered you can largely forget the credentials side and just open chats, which is exactly the low-friction experience a flat-rate plan is meant to give. With the key settled, the experience becomes opening chats and nothing more — which is the point. There is nothing per-session to manage or rotate, and no usage dashboard to keep an eye on.

API Host at the root

Set the API Host to the host root with no /v1 — Cherry Studio appends the version path itself, and adding /v1 yourself doubles it and fails. This is the single most common setup error and it is unrelated to your plan.

If the gateway needs an exact path, end the host with a # to lock it. For most unlimited setups you leave the host bare and let Cherry Studio append.

A trailing slash on the host breaks the path the same way an extra /v1 does, so enter the bare host and nothing more. Getting this one field right is usually the difference between the provider working immediately and chasing a confusing 404 that has nothing to do with your unlimited key. Get this field right and the provider simply works, with no usage-related troubleshooting to do. A bare host is all it ever needs, and the # override is there for the rare gateway with a non-standard path.

Chat without rationing

On metered pricing a long Cherry Studio session quietly adds up, so people self-censor and keep conversations short. On a flat-rate unlimited plan that pressure is gone — you let threads run long and ask follow-ups freely.

Keep several conversations open across the day if you like. Every message draws from the same flat-rate plan, so the bill does not move with how much you chat.

Long threads are where this matters most, because Cherry Studio resends the conversation history with each message, so a deep chat costs progressively more per turn on a meter. Unlimited flattens that entirely — a hundred-message thread and a two-message one sit on the same fixed plan, so you never trim history to save money. Deep threads are where unlimited quietly saves the most, since history resends grow with every turn. On a flat rate a long thread costs no more than a short one, so you never trim context to save money.

Use the strongest model

Add claude-sonnet-4-5 for everyday chat and claude-opus-4-8 for the hardest questions. On unlimited you can default to the stronger model rather than reserving it for special cases to save money.

One unlimited key covers every model and every conversation in Cherry Studio for the plan's duration, so model choice becomes purely about quality, not cost.

That removes a small but constant decision people make on metered access, where every switch to the stronger model is weighed against its price. With unlimited you simply leave the best model selected and let it answer everything, which is both simpler and produces better results. Defaulting to the strongest model is both simpler and better once cost stops being part of the choice. You leave Opus selected and let it answer everything.

# Cherry Studio -> Settings -> Model Services -> + Add
# Provider type: Anthropic   (native /v1/messages)
# API Key:  <your gateway key>
# API Host: https://claudeapikey.dev        (root only, NO /v1 -- it auto-appends)
#   end with #  ->  https://claudeapikey.dev#   to lock the exact URL
# Model: claude-sonnet-4-5

FAQ

無制限 Claude で Cherry Studio を動かせますか？
はい——Anthropic タイプのプロバイダを追加し、無制限キーを貼り付け、API Host を {SITE}（ルート、/v1 なし）に設定します。定額プランがすべてのチャットをカバーします。

なぜ Cherry Studio に無制限が向いているのですか？
長い探索的チャットは着実にトークンを消費します。定額・無制限なら、どれだけチャットしても料金は固定です。

API Host に /v1 を含めるべきですか？
いいえ——ホストのルートのみ。Cherry Studio が自動で付加します。自分で /v1 を付けるとパスが二重になり失敗します。

無制限プランに制限はありますか？
プラン期間中はトークン上限なし。ゲートウェイを速く保つためのフェアユースのレート制限のみです。