Menu

Continue.dev Integration

Continue.dev is an open-source AI code assistant for VS Code and JetBrains IDEs. This guide shows how to connect Snipara to Continue for documentation-aware coding assistance using Server-Sent Events (SSE) transport.

Primary path
For LLM agents: use Hosted MCP
Continue is an LLM agent client. Use Hosted MCP, with SSE only because Continue's transport model requires it.

Recommended Setup

Continue is best configured through direct SSE MCP settings with a project slug and API key. That matches Continue's transport model and keeps the setup explicit.

The local snipara-mcp-login flow is separate compatibility tooling and is not the default model shown on this page.

Transport Matrix

Use the project slug consistently across all MCP transports. Pick the transport your client actually supports instead of mixing setup stories.

Project slug
HTTP direct
Available

Best when the client accepts plain remote MCP over HTTP and you want the simplest setup.

https://api.snipara.com/mcp/YOUR_PROJECT_SLUG
SSE
Recommended

Use only for clients that explicitly require streaming Server-Sent Events transport.

https://api.snipara.com/mcp/YOUR_PROJECT_SLUG/sse
Local stdio
Local only

Use for stdio compatibility, client package development, and advanced tests powered by uvx snipara-mcp. Prefer Hosted MCP when remote MCP is supported.

SNIPARA_PROJECT_SLUG=YOUR_PROJECT_SLUG

What You Get

MCP Tools in Chat

Access Snipara tools directly from Continue's chat interface using @snipara.

Open Source

Continue is fully open source. Use with any LLM provider (OpenAI, Anthropic, local models).

SSE Streaming

Real-time streaming responses via Server-Sent Events for fast, interactive conversations.

Prerequisites

  • Continue.dev extension installed in VS Code or JetBrains IDE
  • A Snipara project with at least one document indexed
  • A project slug and a Snipara API key with access to that project

Setup

Step 1: Open Continue Configuration

In VS Code, open the Command Palette (Cmd/Ctrl + Shift + P) and search for "Continue: Open Config". This will open your ~/.continue/config.json file.

Step 2: Add MCP Server Configuration

Add the Snipara MCP server to your configuration using SSE transport:

Direct SSE Path

Continue does not use the local snipara-mcp login flow here. This config is a direct remote MCP setup, so auth is placed in the Continue configuration itself.

{
  "models": [
    "..."
  ],
  "mcpServers": [
    {
      "name": "snipara",
      "transport": {
        "type": "sse",
        "url": "https://api.snipara.com/mcp/YOUR_PROJECT_SLUG/sse",
        "headers": {
          "X-API-Key": "YOUR_API_KEY"
        }
      }
    }
  ]
}

SSE Transport

Continue.dev uses Server-Sent Events (SSE) for streaming responses from MCP servers. The SSE endpoint is /mcp/YOUR_PROJECT_SLUG/sse on the hosted MCP origin (note the /sse suffix).

Step 3: Replace Placeholders

Replace the following values:

  • YOUR_PROJECT_SLUG - Your Snipara project slug
  • YOUR_API_KEY - Your Snipara API key (starts with snp-)

Step 4: Reload Continue

After saving the configuration, reload Continue by running "Continue: Reload Configuration" from the Command Palette.

Using Snipara in Continue

Once connected, you can use Snipara tools in Continue's chat interface:

@snipara How does the authentication system work?
[Using snipara_context_query with max_tokens=4000]
Based on your documentation, the auth system uses...

Available Tools

The following Snipara tools are available through Continue:

Auth Model Reminder

This page is the API-key-based Continue integration. Use local snipara-mcp plus snipara-mcp-login only for stdio compatibility, client development, or advanced testing.

snipara_context_query

Query documentation with token optimization

snipara_ask

Simple question-answer queries

snipara_search

Search for patterns (regex supported)

snipara_remember

Store facts, decisions, learnings

snipara_recall

Semantically recall stored memories

snipara_decompose

Break complex questions into sub-queries

snipara_multi_query

Execute multiple queries in one call

See the MCP Tools Reference for complete documentation of the full 125-tool contract.

Code Graph v1

In v1, snipara_context_query stays doc-first. It remains the right default for docs, notes, markdown, and narrative context.

For structural code questions, use snipara_code_callers, snipara_code_imports, snipara_code_neighbors, or snipara_code_shortest_path.

Workflow Examples

Query Documentation

@snipara How does the rate limiting work?
[Using snipara_context_query with max_tokens=4000]
Based on your documentation, rate limiting is implemented...

Search for Code Patterns

@snipara Search for usages of useAuth hook
[Using snipara_search with pattern="useAuth"]
Found 12 matches in 5 files...

Memory-Driven Development

Save a decision for later
@snipara Remember: Using JWT for auth tokens because...
Memory saved with type=decision
Recall later in a new session
@snipara What did we decide about authentication?
[Using snipara_recall with query="authentication"]
Found decision: Using JWT for auth tokens because...

Alternative: HTTP Transport

If your version of Continue doesn't support SSE, you can use the standard HTTP transport:

{
  "mcpServers": [
    {
      "name": "snipara",
      "transport": {
        "type": "http",
        "url": "https://api.snipara.com/mcp/YOUR_PROJECT_SLUG",
        "headers": {
          "X-API-Key": "YOUR_API_KEY"
        }
      }
    }
  ]
}

Troubleshooting

Connection timeout or failure
  • Verify your API key is correct and not expired
  • Check that the project slug in the URL is correct
  • Ensure you're using the project-scoped SSE endpoint (/mcp/YOUR_PROJECT_SLUG/sse) rather than the non-streaming HTTP endpoint
  • Try the HTTP transport as an alternative
  • If SSE is not available in your environment, use local snipara-mcp and snipara-mcp-login as a compatibility path.
Tools not showing in Continue
  • Reload Continue configuration after saving changes
  • Check the Continue output panel for error messages
  • Ensure the mcpServers array is properly formatted JSON
  • Try restarting VS Code/JetBrains
"Rate limit exceeded" error
  • Check your usage in the Snipara dashboard
  • Upgrade your plan for higher limits
  • Wait a minute for the rate limit to reset
  • Trim the query if you are asking Continue to fetch too much context at once
SSE connection drops frequently
  • Check your network connection stability
  • If using a VPN or proxy, try without it
  • Consider using HTTP transport instead for more stable connections

Next Steps