Continuum Context Hub
The universal AI memory layer — seamlessly switch between AI coding assistants without losing your project context.
Main Features
- Store and recall project knowledge (architecture decisions, coding conventions, debugging notes) as persistent memories
- Cross-tool sync — memories saved in VS Code are instantly available in Cursor, Windsurf, Antigravity, and any Continuum-connected IDE
- Organize memories by category (architecture, conventions, patterns, debugging, decisions, preferences) and importance (critical → ephemeral)
- Semantic search — recall the right context using natural language queries
- Sidebar tree view — browse all project memories at a glance, grouped by category with importance indicators
- Managed server — the extension automatically installs and runs the Continuum server locally
- Works with the MCP protocol for Claude Code and other AI agent frameworks
Usage
Store a Memory
- Open the Command Palette (
Cmd+Shift+P / Ctrl+Shift+P)
- Run Continuum: Remember
- Type what you want to remember — e.g., "We use camelCase for all TypeScript functions"
- Pick a category and importance level
Recall a Memory
- Run Continuum: Recall from the Command Palette
- Type a natural language query — e.g., "naming conventions"
- Results are ranked by semantic relevance, importance, and freshness
Push Selection
Select any block of code or text → run Continuum: Push Selection to store it as a memory with full context.
How It Works
VS Code ──┐
Cursor ──┤──▶ Continuum Server ──▶ Shared Memories
Windsurf ──┘ (runs locally) (per-project)
The Continuum server runs on your machine and stores memories in a local SQLite database with optional vector embeddings (ChromaDB + Sentence Transformers) for semantic search. No data leaves your machine unless you configure a remote server.
Commands
| Command |
Description |
Continuum: Connect |
Verify connection to the server |
Continuum: Remember |
Store a new memory |
Continuum: Recall |
Semantic search across memories |
Continuum: Project Context |
View all memories as a formatted document |
Continuum: Sync |
Refresh project context |
Continuum: Push Selection |
Store selected text as a memory |
Continuum: Delete Memory |
Remove a memory (sidebar right-click) |
Continuum: Copy Memory Content |
Copy to clipboard (sidebar right-click) |
Settings
| Setting |
Default |
Description |
continuum.serverUrl |
http://localhost:8111 |
Continuum server URL |
continuum.serverPort |
8111 |
Port for the managed server |
continuum.serverManaged |
true |
Auto-start and manage the server |
continuum.autoPushOnSave |
false |
Push a memory on every file save |
The Continuum panel in the Activity Bar displays your project memories grouped by category, with importance indicators:
| Icon |
Importance |
| 🔴 Error |
Critical |
| ⚠️ Warning |
High |
| ℹ️ Info |
Medium |
| ○ Circle |
Low |
| 🕐 Clock |
Ephemeral |
Right-click any memory to delete or copy its content.
Requirements
- Python 3.10+ — the Continuum server is a Python package
- VS Code 1.80+
Install the server:
pip install continuum-context-hub[embeddings]
Or let the extension install it for you on first launch.
Installation
VS Code Marketplace:
ext install FormulateAI.continuum
Open VSX (VSCodium / Gitpod):
Available at open-vsx.org
From Source:
git clone https://github.com/FormulateAI/continuum.git
cd continuum
npm install && npm run compile
# Press F5 in VS Code to launch in development mode
Telemetry
This extension collects anonymous usage data to help improve the product:
- Extension activation and command usage (which features are popular)
- Error reports (what's failing)
- VS Code version and OS type
No personal data, code, or file contents are ever collected.
Telemetry respects your VS Code setting: set telemetry.telemetryLevel to "off" to disable all data collection.
License
Apache 2.0
