Cursor Integration
Cursor is an AI-powered code editor built on VS Code. Snipara integrates with Cursor through two complementary channels: a native extension (via OpenVSX Registry) for UI and commands, and an MCP server for AI chat access. Use both for the best experience.
OpenVSX Registry
Cursor uses the Open VSX Registry for its extension marketplace — an open-source, vendor-neutral alternative to the VS Code Marketplace. The Snipara extension is published to both registries, so you get the same features in Cursor as in VS Code.
Direct URL: cursor:extension/snipara.snipara
What You Get
43 Commands
Full coverage of all Snipara MCP tools plus RLM Runtime — query, search, memory, swarms, documents, runtime execution, and more.
4 Sidebar Views
Dedicated activity bar icon with Results, Session Context, Memories, and Swarm Dashboard panels.
- ✓ Results (TreeView)
- ✓ Session Context
- ✓ Memories (TreeView)
- ✓ Swarm Dashboard (WebView)
MCP AI Chat
Cursor's AI can call Snipara tools directly in chat and composer mode via MCP integration.
- ✓ Query Documentation
- ✓ Store & Recall Memory
- ✓ Search & Plan
- ✓ Team Standards
About OpenVSX Registry
The Open VSX Registry is an open-source extension registry operated by the Eclipse Foundation. It serves as the primary extension marketplace for:
- Cursor — The AI-powered code editor
- VSCodium — The free/libre VS Code distribution
- Gitpod — Cloud development environments
- Eclipse Theia — The open-source IDE framework
- Code-OSS — The open-source VS Code base
Why OpenVSX?
Microsoft's VS Code Marketplace has licensing restrictions that prevent non-Microsoft products from using it. OpenVSX provides the same extension ecosystem without these restrictions. Snipara publishes to both registries simultaneously, so you always get the latest version regardless of your editor.
Install Extension (Recommended)
Option A: From Cursor Extensions (Easiest)
- Open Cursor
- Go to Extensions (Cmd+Shift+X / Ctrl+Shift+X)
- Search for "Snipara"
- Click Install
Cursor automatically fetches extensions from the OpenVSX Registry.
Direct URL:
cursor:extension/snipara.sniparaOption B: Direct from OpenVSX
- Visit open-vsx.org/extension/snipara/snipara
- Click Download to get the
.vsixfile - In Cursor: Extensions →
...menu → Install from VSIX
Option C: From GitHub Releases
Download the latest .vsix from GitHub Releases, then install via the Extensions menu.
Extension Features in Cursor
| Feature | Cursor Support | Notes |
|---|---|---|
| 43 Commands (Command Palette) | ✓ Full | Type "Snipara:" in palette |
| 4 Sidebar Views | ✓ Full | Results, Context, Memories, Swarm |
| Keyboard Shortcut (Cmd+Shift+R) | ✓ Full | Quick documentation query |
| File Explorer Context Menu | ✓ Full | Upload .md/.mdx/.txt files |
| Status Bar Item | ✓ Full | Click to query documentation |
| Configuration Settings | ✓ Full | snipara.* in Cursor Settings |
| GitHub Copilot Language Model Tools | ✗ N/A | Copilot-specific API, use MCP instead |
| MCP Server Definition Provider | ✗ N/A | Copilot-specific, use MCP config instead |
Command Palette access
Results, Context, Memories, Swarm
Cmd+Shift+R
Upload documents
Click to query
snipara.* in settings
Use MCP integration instead
Cursor vs VS Code
All extension features work in Cursor except GitHub Copilot-specific integrations (Language Model Tools and MCP Server Definition Provider). For AI chat access to Snipara tools, use the MCP integration below alongside the extension.
Configuration
Quick Setup
- Open Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
- Run "Snipara: Configure"
- Enter your API key from the Snipara Dashboard
- Enter your project ID (or slug)
Or configure directly in Cursor Settings (settings.json):
{ "snipara.apiKey": "rlm_your_api_key", "snipara.projectId": "your-project-slug", "snipara.serverUrl": "https://api.snipara.com", "snipara.maxTokens": 4000, "snipara.searchMode": "hybrid"}MCP Integration (AI Chat Access)
Configure Snipara as an MCP server so Cursor's AI can call Snipara tools directly in chat and composer mode. This complements the extension by giving Cursor's AI access to your documentation.
Prerequisites
- Cursor version 0.40 or later (MCP support required)
- A Snipara project with at least one document indexed
- An API key from your Snipara dashboard
MCP Setup
Step 1: Open MCP Settings
In Cursor, go to Settings → Features → MCP Servers.
Step 2: Add Configuration
Click Edit Config and add the following to your mcp_config.json:
{ "mcpServers": { "snipara": { "command": "npx", "args": [ "-y", "@anthropic-ai/mcp-proxy", "https://api.snipara.com/mcp/PROJECT_SLUG" ], "env": { "MCP_API_KEY": "YOUR_API_KEY" } } }}MCP Proxy
This uses the @anthropic-ai/mcp-proxy package to connect HTTP MCP servers. It will be installed automatically via npx on first use.
Step 3: Enable the Server
After saving the configuration:
- Return to the MCP Servers settings
- Find "snipara" in the list
- Toggle it ON
- Wait for the connection indicator to turn green
Using Snipara in Cursor Chat
Once connected, you can use Snipara tools in Cursor's AI chat (Cmd/Ctrl + L):
How does the authentication system work in this project?[Using rlm_context_query]Based on your documentation, the auth system uses JWT...Composer Mode
Snipara works great with Cursor's Composer mode (Cmd/Ctrl + I). When editing code, Cursor will automatically query your documentation for context.
Available MCP Tools
With MCP integration, Cursor's AI has access to all 43 Snipara tools:
| Category | Tools |
|---|---|
| Core Query | rlm_context_query, rlm_ask, rlm_search, rlm_multi_query |
| Planning | rlm_decompose, rlm_plan, rlm_multi_project_query |
| Agent Memory | rlm_remember, rlm_recall, rlm_memories, rlm_forget |
| Swarm Coordination | rlm_swarm_create, rlm_swarm_join, rlm_claim, rlm_task_* |
| Team Context | rlm_shared_context, rlm_list_templates, rlm_get_template |
| Documents | rlm_upload_document, rlm_sync_documents, rlm_store_summary |
Alternative: Direct HTTP
If Cursor supports direct HTTP MCP servers in your version, you can use this simpler configuration:
{ "mcpServers": { "snipara": { "transport": "http", "url": "https://api.snipara.com/mcp/PROJECT_SLUG", "headers": { "X-API-Key": "YOUR_API_KEY" } } }}Workflow Examples
Quick Documentation Query (Extension)
Press Cmd+Shift+R (or Ctrl+Shift+R)How does authentication work?[Snipara] Querying documentation...Found 3 relevant sections (2,847 tokens)Results appear in the Snipara sidebarAI Chat Query (MCP)
In Cursor Chat (Cmd+L):How is rate limiting implemented in this project?[Using rlm_context_query]Based on your documentation, rate limiting uses...Memory-Driven Development
Save a decision via Command Palette:Snipara: RememberType: decisionContent: "Chose JWT over sessions for auth"Memory stored. View in Memories panel.Later, in AI chat:What authentication approach did we decide on?[Using rlm_recall]You decided to use JWT over sessions for auth...Troubleshooting
Run Snipara: Configure from the Command Palette and enter both your API key and project ID. Or set snipara.apiKey and snipara.projectId in Cursor settings.
Check that the URL and API key are correct. Verify @anthropic-ai/mcp-proxy is working with npx @anthropic-ai/mcp-proxy --version. Try restarting Cursor or check the developer console for error messages.
Ensure the MCP server toggle is ON (green indicator). Try explicitly asking: "Use rlm_context_query to find...". Check if there are any error messages in the MCP settings panel.
Make sure Node.js is installed and in your PATH. Try using the full path: /usr/local/bin/npx. Restart Cursor after installing Node.js.
Make sure you're searching for "Snipara" (not "RLM"). Cursor uses the OpenVSX Registry — try downloading directly from open-vsx.org and installing via VSIX.
Recommended: Use Both
For the best Cursor experience, install both the extension and the MCP integration:
| Channel | What It Provides | Best For |
|---|---|---|
| Extension (OpenVSX) | Sidebar views, 43 commands, keyboard shortcuts, file explorer integration, status bar | Manual queries, browsing results, memory management |
| MCP (AI Chat) | AI chat access — Cursor's AI can query your docs in chat and composer mode | Conversational queries, code generation with context |
Extension (OpenVSX)
Sidebar views, 43 commands, keyboard shortcuts, file explorer integration, status bar
Best for: Manual queries, browsing results, memory management
MCP (AI Chat)
AI chat access — Cursor's AI can query your docs in chat and composer mode
Best for: Conversational queries, code generation with context