Skip to content
llmwise
Sign inStart free

Help

llmwise in Claude

Add llmwise to Claude, and Claude can ask the other models on your llmwise plan for you: one for a second opinion, or two or three side by side. You sign in with your llmwise account, and each answer counts toward your plan's messages, as on llmwise.ai.

Start freeSee pricing

What llmwise does in Claude

llmwise adds three tools to Claude. Claude picks one from what you ask, and shows each model's answer in its own column.

  • Ask a model

    Asks one model a question and shows its answer. For a second opinion from a different model, or when you name the model you want.

  • Compare models

    Asks two or three models the same question at once and shows their answers side by side.

  • List models

    Lists the models your llmwise account can use, with who makes each one. It only reads, so it uses no messages.

Add llmwise to Claude

From Claude's connector directory:

  1. In Claude, open Customize → Connectors and choose Discover.

  2. Search for llmwise and open it.

  3. Choose Connect to Claude, then sign in as below.

If llmwise isn't in the directory for you yet, add it as a custom connector instead:

  1. In Claude, open Customize → Connectors and choose Add custom connector.

  2. Name it llmwise, and enter https://llmwise.ai/mcp as its URL. Leave the OAuth fields empty.

  3. Choose Add, then Connect, and sign in as below.

On Claude's Free plan you can add one custom connector. On a Team or Enterprise plan, an Owner adds it for the organization.

Sign in and turn it on

  1. Claude opens llmwise's sign-in page. Choose Continue with email, and open the link we email you. New to llmwise? The same link creates your account.

  2. Back on the page, check it shows your email and choose Connect, then approve the access Claude asks for. You're back in Claude, connected: it can see the models your account can use and ask them questions for you.

  3. In a chat, select + → Connectors and make sure llmwise is on. Claude may ask before it uses a tool: choose Allow once, or Always allow.

Things to ask

  • Ask GPT-6.1 Sol on llmwise for a second opinion on this plan.

  • Compare Gemini 3.1 Pro, Grok 4.7 and DeepSeek V4 Pro on llmwise: which sorting algorithm works best on nearly sorted data?

  • Which models can I use on llmwise?

  • Have two other models on llmwise check this proof, and tell me where they disagree with you.

Messages, and what's kept

Each model that answers counts as one message on your llmwise plan, as a message you send on llmwise.ai does: comparing 3 models counts 3. On the free trial, that's 5 messages, once. A model that fails, or takes too long and is stopped after 50 seconds, isn't counted.

The model you ask sees only the question Claude writes for it, not your Claude conversation. A question can be up to 32,000 characters.

Questions and answers from Claude aren't saved to your llmwise chats. Each one leaves a usage record, as every message does: the model, the tokens, the cost and the time, never the text. The Privacy Policy has the details.

Disconnect

In Claude, open Customize → Connectors, select llmwise and choose Disconnect. Claude can no longer ask llmwise anything until you connect again; Remove, in the same page's menu, takes it off your list too. Deleting your llmwise account also ends it.

Help

Something not working, or a question about your account? Email hello@llmwise.ai.

Questions

Does the other model see my Claude conversation?

No. Claude writes a question for the model and llmwise sends only that, so the model doesn't see your chat, your files or your memory in Claude. Claude decides what goes into the question, so it may include parts of the conversation it needs.

Do I need a paid llmwise plan?

You need an llmwise account. The free trial's 5 messages work here too, on the models it includes; after that, a paid plan's messages do. List models shows what your account can use.

Why does Claude ask me before it uses llmwise?

Claude can ask before it runs a connector's tool, unless you've chosen Always allow for it. Ask a model and Compare models spend your llmwise messages, so you may want to keep that check; List models only reads, so Claude may run it without asking. You can change each tool's setting under Tool permissions, on llmwise's page in Customize → Connectors.

Claude, GPT, Gemini, DeepSeek, Grok, Kimi, GLM, and Mistral, in one chat.

See what a message costs before you send it. Free is 5 messages to try; sign in with an email link, no password or card.

Start freeSee pricing