How belleq works.

Introduction

Modern AI assistants are most useful when they can reach your real tools — your docs, your issues, your designs. But wiring each assistant to each tool is repetitive, hard to scope, and locks your setup to whichever AI you used last.

belleq is a single layer that holds your connections, your skills, and your memory. You connect a tool once, add skills from a curated library, organize them into contexts, and hand any AI a single link. The AI sees exactly the tools and skills that context allows — and belleq quietly remembers what matters so it's there next time, on any assistant.

Quickstart

Four steps from zero to an AI working inside the right context.

Core concepts

A handful of ideas cover everything belleq does.

Connect a tool

Open Connectors in the dashboard and choose a tool. You'll sign in to that tool once through its own secure login — belleq never sees your password, and stores the connection encrypted. From then on the connection is available to every context you create; there's nothing to re-enter.

• Connections are unlimited on every plan.
• A connection is shared across contexts — connect Notion once, use it anywhere.
• Revoke a connection at any time; every context that used it updates instantly.

Skills

Connectors give an AI your tools. Skills give it new abilities — ready-made know-how for getting a specific kind of job done. belleq keeps a curated skill library: we review the best skills from across the ecosystem, and you add the ones you want to a context, just like a connection.

Because a skill lives in the context — not in one assistant — it travels with everything else. The same skill is there whether you work in Claude today, ChatGPT tomorrow, or leave an agent to run on its own.

Create a context

A context is a named set of tools for one job. Create one for design, another for support, another for code-review. For each, pick which of your connections and skills it's allowed to use. The AI pointed at that context will see those tools and skills — and nothing else in your workspace.

Every context gets its own stable link in the form:

Connect your AI

belleq speaks the Model Context Protocol (MCP), so it works with any assistant that supports remote connectors. Add a context's link as a custom connector and authorize once.

Knowledge base

As you work, belleq watches the exchange between you and your AI — the documents you share, the answers it looks up, the decisions you reach — and captures what looks durable into the context's knowledge base. There's nothing to tag and nothing to maintain.

When a later conversation touches the same ground, belleq surfaces the relevant knowledge automatically. Important things stay close; stale things fade. Because the knowledge lives in the context, it's there whether you return on Claude, ChatGPT, or anything else.

Agents

A context doesn't just wait to be used — it can work on its own. An agent is an autonomous worker that lives inside a context and already knows its tools and its knowledge base. You give it a role and a task in plain language, and it carries the task out — reading and writing knowledge and calling connectors exactly like you would.

Create an agent from the Agents page in the dashboard, pick the context it belongs to, and describe its job. Run a task straight away to watch it work step by step, or schedule it to run on its own.

Agents run on belleq's managed model out of the box, or you can bring your own key (BYOK or OpenRouter). Either way, a daily budget caps spend, and writes an agent marks as shared wait in a review queue for your approval before they enter the knowledge base.

Access & sharing

A context is a scope. Share one with a teammate and they get exactly those tools — never the rest of your workspace. Each context keeps its own memory and a clear record of what the AI saw and used, so access stays accountable.

FAQ