Database-free · runs locally · open source

Your contexts, ready for any AI.

Assemble Confluence pages, Jira issues, repo files and notes into reusable contexts. Preview them inline, then serve any context to Claude, Cursor or your IDE over MCP.

A quick tour: build a context, preview it inline, and hand it to your AI over MCP.

Features

Everything around contexts — nothing you don't need

Contexts

Group repo files, tickets, docs and instructions into reusable, shareable contexts — with mixins and a live preview.

Confluence & Jira

Pull Confluence pages and Jira issues straight into a context. Pages are rendered to clean, LLM-friendly markdown.

Inline file viewer

Browse scratch and repo files in-app — markdown, images, and PlantUML diagrams rendered right there.

Voice input

Dictate notes and instructions with built-in Whisper speech-to-text, with optional AI clean-up.

AI assistant

An assistant on every context proposes items, tags and structure from a plain-language description.

MCP server

Every context is exposed over MCP, so Claude, Cursor or your IDE can read it with first-class tools.

How it works

From scattered docs to AI-ready context in three steps

1

Build a context

Add Confluence pages, Jira issues, repo files, URLs and notes — or let the assistant draft it for you.

2

Preview inline

See the fully rendered context, browse scratch files and PlantUML diagrams, and refine until it's right.

3

Hand it to your AI

Point Claude, Cursor or your IDE at the MCP server and pull the context with one tool call.

No database
Contexts are plain JSON files on disk.
Runs locally
Your code and docs never leave your machine.
Encrypted secrets
API keys encrypted at rest with your password.

Your context, anywhere you code.

Self-hosted, open source, and ready in minutes. Bring your own AI provider.