consens.io

What should the models cross-check?

One answer with its differences and checks.

Send Feedback

Settings

Memory

What consens.io remembers about you across chats. You decide what goes in, and you can change or delete it any time.

Saved memories

Short facts and preferences, saved by Agent or by you. They go with your questions while Memory is on.

    Nothing saved yet.

    About you

    Short, reusable preferences for every question.

    Memory note

    Paste a longer memory summary or keep your own background note here. It is plain text; Luna may add or patch one passage only when you explicitly submit a Remember or Correct memory action.

    When Memory is on, this note is sent with your question to every selected model. It is stored on your account and nowhere in this browser.
    How it works

    It shapes how you are answered, never what is true — where your memory and the evidence disagree, the evidence wins. It is stored on your account, reaches every device, and is deleted with your account. Watch reruns never use it.

    Nothing is collected automatically unless you switch on Let Agent update memory. Then Agent decides what is worth keeping from your own messages — never from web pages, files or emails — shows each change while it works on the answer and lists every saved memory here, where you can edit, delete or undo it. It does not store passwords, keys or account numbers, and saves health, beliefs or similar sensitive details only when you explicitly ask. Switch it off and nothing new is saved; what was saved stays until you delete it here.

    The four About you fields are intentionally short. The server applies your plan’s limit (Free, Plus or Pro) to Saved memories, but a long shared note can pull all six models towards the same framing and make the agreement score look higher than it is. Please leave out health details, beliefs, or other people's personal data.

    Model behavior

    The shared instruction sent to every selected model. It applies to the question; your memory applies to you.

    Customize how the models respond. Leave empty to use the default.

    Runs

    How consens answers your next message.

    Display

    How the app and its results look.

    Appearance

    Switch between light and dark mode.

    Connections

    Use the included allowance or connect your own OpenRouter API key.

    OpenRouter API key

    Manage and test your OpenRouter credential.

    ▼

    Account

    Manage permanent account actions.

    Delete account

    Permanently delete your account, bookmarks, and stored data. This cannot be undone.

    build bed0a62

    FAQ & Quick Start

    What is consens.io?
    consens.io runs your query in parallel across multiple large language models and presents their answers side by side. In addition, it can integrate external search systems to retrieve current and verifiable information.

    Benefits:
    ✓ Systematic comparison of Models
    ✓ Structured consensus building
    ✓ Integration of web-based research sources

    1. What is Auto Consensus? ＋

    Auto Consensus automatically analyses several model responses, identifies agreements and discrepancies, and generates a consolidated answer. The goal is a transparent and robust synthesis of the available outputs.

    2. What does Consensus Button do? ＋

    The Consensus Button combines the selected model responses into a single joint answer. Before running it, you can include or exclude individual responses so that only the models you consider relevant contribute to the final result.

    3. What is the Consensus Model? ＋

    The Consensus Model is a dedicated model that does not answer user queries directly. Instead, it evaluates, structures, and synthesises the already generated model responses into a coherent overall statement.

    4. What does “Differences” show? ＋

    Differences highlights substantive deviations between the model responses. This makes it easier to see where models arrive at different conclusions or add distinct perspectives.

    5. What does the Reasoning switch do? ＋

    With Reasoning on, the same models you selected think longer before they answer. That can help with harder questions, but answers take longer and a run uses more of your daily allowance. It is available on every plan; in Agent mode the same control lets you pick the reasoning level of the chat model.

    6. What does Web Search include? ＋

    Web Search forwards your query to a grounding Model in the background. The crawled webpages are passed to all the models. The results are displayed directly within consens.io.

    7. What is the purpose of including or excluding answers? ＋

    By including or excluding individual model answers, you control which outputs are used in the consensus generation. This allows you to filter out low-quality or irrelevant responses before they influence the final result.

    8. When do I need to enter API keys? ＋

    API keys are only required if you want to use your own model access. You need a verified login before saving, testing, or using personal credentials.

    9. Which AI models are available? ＋

    consens.io provides access to current large language models from several providers, including OpenAI, Mistral, Anthropic, Grok, DeepSeek, Google Gemini, Moonshot AI/Kimi, Z.ai/GLM and Meta Muse.

    10. How many queries can I make? ＋

    Your remaining allowance is shown in the sidebar. Compare, Consensus, Reasoning runs and Agent share it. Usage is limited during early access, and limits may change as the service develops.

    11. What does consens.io cost? ＋

    consens.io is currently free during early access. Usage limits apply, and some features are available only to selected accounts. Paid plans are coming. There is no subscription to buy yet. For more runs, you can connect your own OpenRouter API key in Settings and pay OpenRouter directly. Feedback is welcome at contact@consens.io.