← Back to Documentation

Connect DeerDawn in under 5 minutes

The fastest path to first value is simple: create your workspace, install DeerDawn MCP in one coding tool, sign in once, and confirm your context shows up immediately.

1

Create your account

Start at deerdawn.com/signup. The free plan is enough to reach first value — no billing required.

  • Sign up and create your workspace.
  • Create your workspace, then open your first project.
  • Optionally add a project description or paste in an existing CLAUDE.md or AGENTS.md to seed your context.
  • No API key needed — you sign in via browser OAuth when the MCP server first connects.
2

Connect DeerDawn to your tools

Add one URL and sign in through your browser — no API keys. The same endpoint works in Claude Code, Cursor, Codex, Claude.ai, and ChatGPT.

Remote MCP URL (recommended)

In your tool, add a custom / remote MCP server, paste this URL, then sign in. Leave any OAuth client fields blank.

https://api.deerdawn.com/api/v1/mcp

Advanced: run it locally (optional) — adds repo-file writes

Claude Code

Run in your terminal — Claude Code has no config file

claude mcp add -s user deerdawn -e DEERDAWN_API_URL=https://api.deerdawn.com -- npx -y @deerdawn/mcp-server@latest

Cursor

File: .cursor/mcp.json or ~/.cursor/mcp.json

{
  "mcpServers": {
    "deerdawn": {
      "command": "npx",
      "args": ["-y", "@deerdawn/mcp-server@latest"],
      "env": {
        "DEERDAWN_API_URL": "https://api.deerdawn.com"
      }
    }
  }
}

Codex

File: ~/.codex/config.toml

[mcp_servers.deerdawn]
command = "npx"
args = ["-y", "@deerdawn/mcp-server@latest"]
startup_timeout_sec = 120

[mcp_servers.deerdawn.env]
DEERDAWN_API_URL = "https://api.deerdawn.com"
3

Verify the connection and import context

After restarting your tool, ask it to run these prompts to confirm DeerDawn is connected and import your current project context:

0. Call start_session (or deerdawn-start_session) — if it returns context, you're already connected; stop here
1. Call get_auth_status (or deerdawn-get_auth_status)
2. Call apply_setup (or deerdawn-apply_setup) for your surface
3. Call import_local_context (or deerdawn-import_local_context) for preview
4. Call import_local_context (or deerdawn-import_local_context) with confirmed=true

Always run start_session first: if it returns context, DeerDawn is already connected (for example via the claude.ai connector) and you can skip the rest — apply_setup and import_local_context won't exist on that build and aren't needed. If your host exposes namespaced MCP tools, use the deerdawn-* variants. If local project files exist (CLAUDE.md, AGENTS.md), run import_local_context preview first, then call it with confirmed=true to complete import. If auth fails, call get_auth_status — it returns a sign-in URL to open in your browser.

4

Confirm your first briefed session

You have reached first value when your tool can read the right workspace context without you re-explaining the project.

  • `list_projects` shows the workspace you just created.
  • `get_context` returns your current task, status, or decisions.
  • Your next prompt can assume DeerDawn already knows the project basics.
  • When a task is finished, run `cleanup_context` with `complete_task=true` so command logs are cleared and the task stays marked completed.
5

Connect web tools

Claude.ai connects via remote MCP — no local install needed. Go to Settings → Connectors → Add custom connector, set the URL to https://api.deerdawn.com/api/v1/mcp, then click Connect and authorize in the browser window that opens. No API key needed.

Web tool setup guide