Continue·MCP server·4 min read

Mails.ai for Continue — drop-in MCP server in VS Code + JetBrains

MCP serveradd to ~/.continue/config.json

Add to ~/.continue/config.json under mcpServers (Continue v1.0+). Restart your IDE.

{
  "mcpServers": {
    "mails": {
      "command": "npx",
      "args": ["-y", "@mailsai/mcp-server"],
      "env": {
        "MAILS_API_KEY": "mk_live_..."
      }
    }
  }
}

Continue is the open-source AI code assistant that runs in both VS Code and JetBrains IDEs. The same MCP server config works across both editor families — one snippet in ~/.continue/config.json and your Continue setup gets send + reply + identity primitives in chat mode and agent mode.

Why Continue + mails.ai

Continue’s value proposition is multi-model + multi-IDE flexibility. The MCP server matches that posture — model-agnostic, IDE-agnostic, runs as a separate process via npx. Common Continue + mails.ai patterns:

Multi-model email drafting. Use Anthropic Sonnet for the customer response draft (high-quality writing), switch to a cheaper local model for classification or batch processing. Same mails.send tool behind both.
Per-IDE agent personas. Define different system prompts and slash commands per workspace. A doc-team workspace agent uses formal voice and emails reviewers; a backend-team workspace agent uses terse voice and emails on-call.
Cross-IDE task continuity. Start a task in JetBrains, finish in VS Code. The MCP server config travels with the user; mails.* tools work identically in both.

Setup

Get an API key. api.mails.ai/keys at Phase 1 launch, or request a closed-beta key from support@mails.ai.
Open ~/.continue/config.json in your editor.
Add the snippet from the install card above. If you already have an mcpServers block, just add mails as a sibling key. Replace mk_live_... with your key.
Restart your IDE.The next Continue session spawns the MCP server via npx on first tool invocation. Verify in Continue’s settings panel that mails appears under MCP servers.

First commands to test

# In Continue chat (VS Code or JetBrains)
> What mails.* tools are available?

# Send a test email
> Use mails.send to email myself a hello message from agent "test".

# Read recent inbound
> Get the last day of threads on the test agent and summarize them.

Common patterns

Slash command for email tasks. Define "/email": "Use mails.* tools to handle email work for this codebase" in Continue config. Invoke with /email in chat to scope a session to email-handling.
Model routing per task type.Use Continue’s per-task model selection to route email-drafting tasks to a high-quality model, batch-classification to a cheaper one. Both call into mails.send identically.
Per-workspace MCP config. Continue supports workspace-specific configs in .continue/config.json (project-level, gitignored). Use per-project mails.ai keys for billing isolation.

Security considerations

Per-key scope. One mails.ai key per Continue agent. Workspace-level configs override user-level for project-specific keys.
Multi-model means multi-cost. Switching models in Continue affects model spend, not mails.ai cost. Mails.ai billing is per-event regardless of which model drove the call.
Inbound safety. When your Continue agent reads inbound via mails.list_threads, the same prompt-injection rule applies: check event.injection_scorebefore acting. The model your task is currently on doesn’t change the threat model — injection works against any LLM.

Compare against the Cline integration for VS-Code- specific setup, or the Claude Code integration for the original MCP example.

FAQ

Questions developers ask after wiring this up.

Does this work with Continue's multi-model setup?

Yes. Continue lets you configure multiple model providers (Anthropic, OpenAI, local Ollama, etc.) and switch between them per-task. The MCP server is model-agnostic — mails.* tools are available regardless of which model is actively driving the conversation. Use Anthropic Sonnet for high-stakes email drafts, switch to a local model for cheap classification tasks, both reach for the same mails.send tool.

Can I use Continue's per-config slash commands with mails.ai?

Yes. Define a custom slash command like '/email-reviewer' in Continue's config that combines a system prompt with mails.* tool access. Useful when you want a dedicated email-handling persona separate from your default coding assistant.

What about Continue's autocomplete mode — does it use mails.ai tools?

No. Autocomplete is single-shot completion based on cursor context; it doesn't use tools. Tools (mails.* included) only fire in chat mode and agent mode, where the model has a tool-calling loop. Autocomplete continues to work normally — completing your code without firing mails.* calls.

Continue supports both VS Code and JetBrains. Is the integration identical?

Yes. The MCP server is JSON-config-driven and runs identically across both. The config file path is the same (~/.continue/config.json — Continue stores config in the same location regardless of IDE). Restart the IDE after editing the config.