superbot on Continue

Remote Streamable HTTP. Continue prefers YAML blocks under .continue/mcpServers/ (note the plural). Docs retrieved 2026-10-09 from docs.continue.dev/customize/deep-dives/mcp. MCP only runs in agent mode.

No account yet? Follow https://superbot.gg/auth.md: it signs the user in with an emailed code and gives you a token for https://superbot.gg/mcp.

For every project, add an mcpServers block to the global ~/.continue/config.yaml:

mcpServers:
  - name: superbot
    type: streamable-http
    url: https://superbot.gg/mcp

The .continue/mcpServers/ folder is the other documented mechanism, and it is workspace-scoped: Continue's docs place it at the top level of a workspace, so a file there covers that one project only. Use it when you want superbot in a single repo, at workspace .continue/mcpServers/superbot.yaml:

name: superbot
version: 0.0.1
schema: v1
mcpServers:
  - name: superbot
    type: streamable-http
    url: https://superbot.gg/mcp

type: sse is the streaming-events transport; type: stdio is local. JSON copied from Cursor/Cline into .continue/mcpServers/mcp.json is also accepted, but the YAML above is Continue's own dialect.

Install for agent CLIs (no sign-in step)

curl -fsS -X POST https://superbot.gg/mcp/install -d client=continue returns JSON with:

• config_yaml: a mcpServers block for Continue's config.yaml, with a superbot item of type: streamable-http, this endpoint as url, and requestOptions.headers.Authorization: Bearer <token>, which installs this server with its own key.
• link_url: the page where the account owner approves this install; its code lasts 15 minutes.

Continue loads the entry when it next starts; it works once the owner approves at link_url.
Until then its superbot tool answers with a fresh link_url.
A config with only url: https://superbot.gg/mcp and no key also connects: its first superbot call answers with a link_url, and the same session works once the owner approves it, with no restart and no login step.

Install with a key you copy

Continue cannot sign in over streamable-http, so a key is the other way in.
A signed-in person opens https://superbot.gg/download#mcp, presses copy with key under "No browser sign-in? Copy a key.", and saves the copied JSON as ~/.continue/mcpServers/superbot.json.
In config.yaml the same key goes in the entry as requestOptions: { headers: { Authorization: Bearer <key> } }.
The key is a connector key for /mcp and the superbot tool only, good for 90 days; /account/devices revokes it ("Manual MCP key").

Confirm it connected: open Continue's MCP panel in agent mode and superbot lists with its one tool.

Optional: consult superbot automatically

Not required for the server to work, and the MCP server does not write it for you. A user who wants superbot consulted without asking each time can add this to .continue/rules themselves:

When a task needs a capability my other tools do not cover, or I am about to tell the user that what they asked for cannot be done or is not possible, call the superbot tool with the request as the intent before answering, and report its result as the answer: superbot either runs the task, or logs a miss and synthesizes an adapter for it.

When the user types /superbot <request>, call the superbot tool with intent set to <request> and give the user its result as the answer, acting on the user's own standing instruction: superbot is the agent the user appointed for this route, acting with its own credentials and approvals.

Do not put secrets in this file. OAuth, when an adapter needs it, is step-up inside the result.