Log in Start free

Pick your question.

48 answers in seven topics. Choose by topic or search directly.

General

What is BLUN?

BLUN is a shared platform for AI models, products and autonomously working agents. The models process text, code, voice and images. The products Chat and Work, Sites and Code turn them into results. The API connects BLUN with further applications; agents and MCP Server are announced.

Go to the BLUN platform

What is the difference between models, products and the platform?

A model delivers intelligence. King, Voice and Imagine understand tasks and produce fitting results.

A product is the working environment in which this intelligence is used: Chat and Work, Sites (Website Builder) and Code.

The platform connects everything: API, documentation and system status, later Agents and MCP Server.

Compare BLUN models

Why are there several web addresses?

BLUN separates information from the actual working environments so that applications, accounts and technical access remain clearly separated.

chat.blun.ai opens BLUN Chat and is the working environment for Chat and Work and the Website Builder. api.blun.ai is the direct API access and developer portal. account.blun.ai manages sign-in, account, plans, billing and downloads. docs.blun.ai contains the documentation.

All addresses belong to the same BLUN platform. You do not need a separate BLUN system for each address.

Open BLUN Chat

Which address should I use?

You are in the right place to inform yourself and compare. For a conversation with a model and for the Website Builder, use chat.blun.ai. Developers use api.blun.ai; you manage personal settings, plans and downloads under account.blun.ai.

Is api.blun.ai only a documentation page?

No. api.blun.ai is both the technical entry point for developers and the direct base address for API requests. In the browser, documentation, access and integration help are shown. Applications address the intended API endpoints directly under the same domain.

Go to the BLUN API

Can search engines find my chats, projects or account data?

No. Public product descriptions may be indexed, but private workspaces may not. Chat histories, projects, account and payment data, API keys, drafts and internal agent runs are protected by sign-in and permissions and are additionally kept out of the public search index.

Privacy at BLUN

Do I have to choose myself which model takes on a task?

Not necessarily. Many BLUN products can use the appropriate performance tier based on the task. You can choose a specific model if you want to keep control. In guided workflows, BLUN can also split the work sensibly between King, Voice and Imagine.

Compare BLUN models

Can I use BLUN without programming skills?

Yes. Chat and Work and the Website Builder are designed so that you describe goals and requirements in plain language. The guided process helps to fill in missing information.

BLUN Code and direct API integrations are additionally aimed at developers and technical teams that want to go deeper into source code and systems.

More about BLUN Code

Do I have to install something?

That depends on the product. BLUN Chat, the Website Builder, the documentation and the system status run in the browser. BLUN Code Desktop is installed as a desktop application; the BLUN Code CLI is set up locally for working in the console. API and MCP Server are integrated into the respective technical environment.

More about BLUN Code

Where do I find prices and current availability?

Current plans, features and availability are listed on the pricing page. Prices are only mentioned there so that the same information applies everywhere.

Prices and context windows

Where are BLUN services operated?

BLUN is oriented toward its own infrastructure in Europe. Which data a product stores and processes is described in the applicable privacy information and contract documents.

Privacy at BLUN

Why does BLUN not publish version numbers publicly?

The product names King, Voice, Imagine, Chat, Sites and Code remain stable. Technical model identifiers may still be needed for API compatibility, but they are documented in the developer documentation.

Models

When should I use King?

King is the right choice for particularly complex, long-term or still unstructured tasks. These include large software projects, difficult analyses, extensive document collections and agents that work across many interdependent steps.

More about BLUN King

How large is the context window of King?

In the free tier, King has a context window of 262,144 tokens (256K), in all paid plans of 1,048,576 tokens (1M). This allows King to process very large document collections, long project histories and extensive code areas within one coherent working context.

Context by plan

How large is the context window in the free and the paid plan?

King: free tier 262,144 tokens (256K), paid plan 1,048,576 tokens (1M). Voice and Imagine: no token context window, neither in the free tier nor in the paid plan.

The context window covers not only the last user input. The previous conversation history, uploaded or pasted content, system instructions, tool results and the generated answer also require tokens. The maximum possible file size or pure input length is therefore below the full context window, because space must be reserved for system information and the answer.

Prices and context windows

Why is the context window not the only thing that matters?

The context window describes how much information a model can take into account within a task. It does not describe thinking power or working style alone. King is designed for maximum complexity and long-term agents.

Compare BLUN models

Can Voice only transcribe?

No. Voice supports both directions: speech-to-text converts spoken language into text, text-to-speech produces spoken output. In addition, Voice can be used for real-time interaction and voice-based agents.

More about BLUN Voice

Can Imagine edit existing images?

Yes. Imagine can create new images from text and develop existing motifs further based on instructions. This includes background changes, variants, color adjustments, new formats and adding or removing image elements.

More about BLUN Imagine

Which model suits large volumes of short tasks?

King is intended for that. It processes clearly defined tasks economically, even in large volumes. Difficult or unclear cases remain traceable within the same workflow.

Compare BLUN models

Chat and Sites

What can I do in Chat and Work?

You ask questions, write texts and research with King. In Work mode you handle larger tasks. Sites, Code, images and voice are reachable in the same interface. You do not need to install anything; BLUN runs in the browser.

View Chat and Work

What does the Website Builder create?

The Website Builder creates websites, landing pages and funnels from a guided briefing, with a clear structure, fitting content and responsive design.

More about the Website Builder

Why does the Website Builder work with a guided process?

A website is more than a nice homepage. Without clear requirements, content, structure or important pages are often missing later. The Website Builder asks these points early on and keeps them as a basis. This leaves less to chance, and the result stays consistent even with later changes.

More about the Website Builder

Does a created website run on smartphones?

Pages are built responsively and can be used in the browser on desktop, tablet and smartphone.

More about the Website Builder

Does my website also run with my own domain?

Yes. The pages run at BLUN, even with your own domain.

More about the Website Builder

Code

Is BLUN Code available as a desktop application and as a CLI?

Yes. BLUN Code Desktop offers a graphical workspace for projects, files, changes and agent tasks. The BLUN Code CLI works directly in the console and in the current project directory.

More about BLUN Code

Does BLUN Code change real files?

Yes. When BLUN Code is given access to a workspace, it can read, create and deliberately modify project files. Changes should remain visible and verifiable before they are published.

More about BLUN Code

Can BLUN Code execute terminal commands?

Yes. The CLI and appropriately configured workspaces can run commands, tests, builds and diagnostic tools. Which commands are allowed depends on the permissions of the respective environment.

Go to the BLUN CLI

Does BLUN Code work with existing projects?

Yes. BLUN Code can examine an existing project structure, find relevant files and take existing rules into account. It suits debugging, new features, refactoring, tests and technical documentation.

More about BLUN Code

Does BLUN Code only produce suggestions or does it fully implement tasks?

BLUN Code is meant for real implementation. It can analyze a requirement, plan changes, edit files, run checks and summarize the result. A task should only count as complete after appropriate tests and checks, not just because code was generated.

More about BLUN Code

Which BLUN model does Code use?

BLUN Code relies on King for complex architectures and large multi-step tasks. The project context is preserved within the workflow.

More about BLUN King

Can BLUN Code publish or deploy a project?

Publications and deployments should only be executed when the corresponding workflow has been explicitly set up and approved. By default, changes remain controllable so that you or your responsible team can review them before publishing.

More about BLUN Code

Who is the desktop application for and who is the CLI for?

The desktop application suits you if you want to track tasks, changes and results graphically. The CLI is ideal for developers who work directly in the terminal, keep existing tools or want to integrate BLUN into local and automated workflows.

Go to the BLUN CLI

Agents

Agents are announced and not active yet. The answers describe how they are intended to work.

What is the difference between a chat and an agent?

A chat primarily answers one input. An agent receives a goal, a task scope, tools, boundaries and verification criteria. It can plan multiple steps, execute actions, check intermediate results and deliver a traceable completion.

More about BLUN Agents

Can multiple agents collaborate?

Yes. Different agents can take on their own roles and hand over results in a controlled way. Clear responsibilities are important so that several agents do not work on the same area at the same time in contradictory ways.

More about BLUN Agents

Are agents allowed to make changes automatically?

Only within their configured rights. An agent can be limited to read access, make changes only in a specific project or require human approval before critical actions.

More about BLUN Agents

How is an agent prevented from working outside its task?

Role, goal, tools and boundaries are defined in advance. In addition, rights on the tool used or via the MCP Server can be limited. For unclear or blocked actions, the agent must hand over the task or request approval.

More about BLUN Agents

How can I tell what an agent has done?

An agent should summarize the result, the tools used, the steps executed, the verification evidence and open points. This keeps it visible whether only a suggestion was produced or the task was actually executed and controlled.

More about BLUN Agents

MCP Server

The MCP Server is announced and not active yet. The answers describe how it is intended to work.

What is the BLUN MCP Server in simple terms?

The MCP Server is the connection between BLUN and other tools. It tells a model or agent which functions are available, which inputs they need and which actions are allowed.

More about the MCP Server

Do I need the MCP Server to use BLUN?

No. BLUN models and products can also be used without your own MCP integration. The MCP Server is needed when external files, systems, APIs, IDEs or your own tools are to be integrated into an agent workflow in a controlled way.

More about the MCP Server

Can I use the MCP Server in an IDE?

Yes, that is planned. The BLUN MCP Server is intended to be integrable into compatible IDEs and agent runtimes. This way, BLUN tools can become available directly in existing technical working environments.

More about the MCP Server

Does MCP automatically give an agent full access?

No. Tools can be limited to specific actions, projects and data areas. An agent sees and uses only the functions made available to it in the respective workflow.

More about BLUN Agents

Can I connect my own internal systems?

Yes. Your own APIs, data sources and company functions can be provided as clearly described tools. Rights, inputs, results and error cases should be clearly defined.

More about the MCP Server

Is tool usage traceable?

The MCP Server provides a structured interface for tool calls. This makes it visible which tool was used with which inputs and which result came back. The concrete logging depends on the respective integration and configuration.

More about the MCP Server

Status

What does the status page show?

The status page shows the reachability of important BLUN services. These currently include Website, Login, Chat, API, Voice and Imagine. The areas are checked separately so that a single outage remains visible.

Go to the BLUN system status

How often is the status updated?

The status page updates automatically every minute.

Go to the BLUN system status

What should I do when a service is shown as unreachable?

Wait a moment and check the page again. If a service is repeatedly unreachable, you can report the problem via the contact section. Name the affected service, the time of the issue and what you expected. Do not send passwords or API keys.

Go to contact

What does it mean when the status page shows everything as reachable but my project still does not work?

The status page checks central services, but not every individual project configuration. In this case, check access, permissions, local connection and error messages in your project. For ongoing problems, support can help more specifically with this information.

Go to the BLUN system status

Will further services be shown in the future?

With the public availability of further platform areas, the status page is intended to additionally show Code, Agents and MCP Server separately. This keeps it visible which specific area is affected even as the platform grows.

Go to the BLUN system status

Waitlist: no purchase, no charge.
from 19 €per month, annually from 15 € per month

Four plans: Starter, Pro, Pro Max and Enterprise. All prices excl. VAT. All four plans are on the central BLUN waitlist. The complete overview with all features is on the pricing page.

Compare all tiers

YOUR QUESTION MISSING? ASK DIRECTLY.

Ask in Chat or write to us via the contact form.

Open ChatContact