v0.1: Foundation

Your team gets caches as the boundary for memory: structured chunks, a navigable tree (Inbox, databases, documents), semantic search scoped to the active cache, and links between entries so related work stays connected.

This release establishes the model the rest of the product builds on.

Knowledge Layer · Second Brain (organization) · Overview (dashboard & caches)

ContextCache exposes an MCP server so assistants in Cursor, Claude Desktop, Claude Code, and other compatible clients can read and write your cache with the same memory the web app uses—no separate silo for "IDE context."

v0.1 includes the integration path, documentation for each client, and the worker endpoint you connect to from those environments.

MCP Server docs

Blueprints are versioned, configurable workflows: instructions, data bindings to your cache and connections, triggers (including chat and schedules), and outputs back into memory or side channels. The dashboard includes a blueprint builder and workspace-level management.

Built-in blueprint definitions ship with the product; the blueprint library in docs will grow as we document each one. You enable and tune blueprints per workspace without deploying code.

Blueprints · Blueprint Library (catalog in progress)

The plugin catalog and connections (OAuth-backed integrations) let blueprints and sync jobs use external services—email, calendar, and more as we expand the catalog.

Plugins · Connections

• Chat in the product against your cache, with system rules and blueprint triggers.
• Hook API for pushing and pulling data with scoped keys (docs).
• System rules (workspace and per-cache instructions) to ground AI behavior (System Rules).
• Billing & plans hooks for teams using the hosted product.
• Documentation — guides for memory, blueprints, MCP, APIs, and how the pieces fit together (/docs).
• Users and Teams — team switcher, membership, and admin flows for who can access which caches (Dashboard → Users & Teams, tenant admins).