com.cambercloud.camber-mcp/camber-mcp-server

Official5 tools

Camber - Remote Agentic Data Science

Remote data science agents for Snowflake, Databricks & BigQuery in Claude/Cursor via MCP

Provides remote data science agents for Snowflake, Databricks, and BigQuery.

Captured live from the server via tools/list.

usage_guide

Instructions for correctly using this MCP server. Call at the start of a session to learn tools, auth, CLI prerequisites, and stash file handling.

No parameters.

agents_create

Create a Camber agent. Creates a personal agent by default. Pass `team_name` to create the agent for a team you belong to. Returns the created agent, including its `tag`.

Parameters (6)

namestring
Name of the agent to create.
aliasstring
Alias used in the agent tag, for example `my-agent` in `@owner.my-agent`.
descriptionany
Optional short description of the agent.
instructionsany
Optional instructions that define the agent behavior.
team_nameany
Optional team unique name. When omitted, creates a personal agent.
structured_outputboolean
Whether the agent should return structured output.

user_info

Get the authenticated Camber user and team names. Returns email, username, and teams list.

No parameters.

agents_chat

Chat with a Camber agent with the specified tag and message. Powered by Camber CLI: `camber chat --agent <agent_tag> --message <message>` For example: `camber chat --agent @owner.agent-alias --message "Hello Camber!"` Returns: The agent's reply text. The run may reference **output files in Camber Stash** (`stash://...` paths). Those files live in the cloud until the user copies them locally. **After the call — sync stash outputs (user's terminal, Camber CLI):** 1. Note every `stash://user/...` path the agent produced or mentioned. 2. Download to the user's machine, e.g. `camber stash cp stash://<user>/<path/to/file> ./<local_filename>` (use a directory path for `./` if copying a folder; see `camber stash cp --help`). 3. To browse first: `camber stash ls stash://<user>/...` 4. Confirm with the user where files were saved. Docs: https://docs.cambercloud.com/docs/stash/ and https://docs.cambercloud.com/docs/camber-cli/stash

Parameters (4)

agent_tagstring
Tag of the agent you want to chat with. Start with @ to use the agent's tag.agent_tag = `@owner_name.agent_alias`For example, @camber_user.ai-agent-1
messagestring
The message to send with the agent.
model_idstring
The name of the model to use for the agent. For example, Haiku, Sonnet
conversation_idany
Optional UUID of an existing conversation to continue. Pass the conversation_id returned by a previous agents_chat call to maintain context across multiple turns.

agents_list

List Camber agents with optional scope and pagination. Returns: AgentListInfo object containing: - data: List of AgentInfo objects - page: PageInfo object with pagination details (page, size, total, num_page)

Parameters (4)

scopeany
Scope of agents to list. Use 'community' for community agents, 'system' for system agents, 'team' for a specific team, or leave empty (null) to list your personal agents.
team_nameany
Unique name of the team whose agents you want to list. Required when scope is 'team'. Ignored for other scopes.
pageinteger
The page number to list agents from (default: 1).
sizeinteger
The number of agents to list per page (default: 20).

Install

claude_desktop_config.json

{
  "mcpServers": {
    "camber-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://camber-mcp.cambercloud.com/mcp"
      ]
    }
  }
}

Desktop config is stdio-only; this bridges via mcp-remote. Native remote: Settings > Connectors.