Continue is the open-source AI coding assistant for VS Code. It speaks MCP natively, so adding ContextBolt is one config change away from full bookmark search inside your editor.
Why this combination works
Most coding sessions involve looking things up. Library docs. Stack Overflow threads. That tweet someone posted about a clever pattern you saved last month. The looking-up pulls you out of flow.
Developers who save technical content on Twitter, Reddit, and LinkedIn end up with a personal reference library. The problem is that this library lives in a browser tab while you’re coding in an editor. Every reference check is a context switch.
The Continue + ContextBolt integration kills that switch. Your saves are queryable from the same sidebar where you’re already chatting with Claude, GPT, or whichever model Continue is configured to use.
The VS Code + Continue MCP setup
Three minutes if Continue is already installed. Open config.yaml (Continue: Open config.yaml in the Command Palette) and add the bookmarks server, swapping YOUR_TOKEN for the token from your extension settings:
mcpServers:
bookmarks:
type: http
url: https://api.contextbolt.app/mcp/YOUR_TOKENSave, reload VS Code, open the Continue sidebar. Ask Continue to search your bookmarks. If the MCP connection is healthy, results appear inline in the chat. We tested this against Continue’s current and previous minor versions; the older one needed the mcp-remote bridge in the FAQ rather than the direct URL field.
VS Code + Continue MCP workflow patterns that earn the setup time
Reference lookups during refactors. “Search my bookmarks for anything about migrating from REST to gRPC.” Continue surfaces saved threads, articles, and Reddit posts on the topic without you opening a browser tab.
Pre-flight reading for new tasks. Before starting a task, ask Continue what you’ve saved about the area. You’ll often find you already collected the right reference material weeks ago.
Combining code context with saves. Continue understands the file you have open. Combine that with bookmark search: “I’m working on the rate limiter in src/middleware/. Do I have any saves about token bucket vs leaky bucket?”
When you wouldn’t use VS Code + Continue with MCP
If you don’t save technical content socially, the integration adds nothing. ContextBolt earns its place when you’re actively saving threads and posts as you browse Twitter or Reddit during downtime.
If you prefer command-line workflows, Claude Code gives you the same access from the terminal. Pick the surface that matches where you actually work.
How to connect ContextBolt to VS Code + Continue
-
Install the Continue extension in VS Code
Open the VS Code Extensions panel and install 'Continue', the open-source AI coding assistant. Continue supports MCP servers natively, so it can talk to ContextBolt directly. If you already use Continue, skip this step.
-
Get your ContextBolt MCP token
Open the ContextBolt extension popup in Chrome, click Settings, and copy your MCP token from the MCP section. You need a Pro subscription ($6/month) for MCP access.
-
Add ContextBolt to Continue's config
Open Continue's config file (Cmd/Ctrl+Shift+P then 'Continue: Open config.yaml'). Under mcpServers, add ContextBolt as a remote HTTP server pointing at https://api.contextbolt.app/mcp/YOUR_TOKEN. Save and reload VS Code. Older Continue versions on config.json need the mcp-remote npm bridge instead, covered in the FAQ.
-
Test the connection
Open the Continue chat sidebar and ask 'Search my bookmarks for anything about API design'. Continue should call the ContextBolt MCP tool and return matching saves. If nothing happens, check the Continue logs panel for connection errors.
Example prompts for ContextBolt + VS Code + Continue
Once connected, try asking VS Code + Continue:
Pull saved technical reference material into a refactor session. Continue searches semantically, so saved threads with different terminology (sessions, tokens, auth flows) all surface.
Reddit threads are often deep technical discussions. Filtering by platform brings back the long-form content rather than tweet-sized takes.
Use ContextBolt as a curated alternative to Stack Overflow. Your bookmarks are filtered for relevance once already, by you saving them.
Time-boxed search. Useful for catching up on a topic you've been bookmarking but haven't sat down to read yet.
What you can do with ContextBolt in VS Code + Continue
Search bookmarks from the Continue sidebar in VS Code
Pull saved threads, articles, and discussions into chat context
Filter by platform (Twitter/X, Reddit, LinkedIn) or keyword
Combine bookmark search with Continue's code understanding
Works in any VS Code-based editor (Cursor, VSCodium, GitHub Codespaces)
Save new bookmarks or search existing ones (no editing or deleting from the editor)