DeepSieve MCP server

Two ways to connect. This package is the stdio + API-key server — ideal for CI, containers, and services. Where a deployment has the remote server with browser login enabled, that route is easier for a developer at their editor — no key to manage. It is not enabled on every deployment, so see https://deepsieve.ai/developers/mcp for the current endpoint rather than assuming one.

Gives any MCP client (Claude Code, Cursor, Claude Desktop, …) outcome-shaped tools over the DeepSieve /v1 API. With them an agent can check the account, discover and propose Blueprints, estimate, start and follow research runs, read and export the cited dataset, review held and conflicting cells, manage re-research and monitoring, work with reports and chat, and record what your team knows about the field. The full list is below.

It is a plain API client — it holds no privileged access, only the scoped key you give it. Mint one with the Agent preset at /settings/api-keys: it can run research, read results and create Blueprints, but it can't change an active Blueprint's schema or your billing.

Install

Pinned to 0.1.0? Upgrade. It doesn't start, because it allowed mcp 2.x, which renamed FastMCP. Later versions cap the dependency at mcp<2, so plain uvx deepsieve-mcp works. If you can't upgrade, run uvx --with 'mcp<2' deepsieve-mcp. The port to mcp 2.x is tracked in #796.

# Claude Code
claude mcp add deepsieve \
  --env DEEPSIEVE_API_KEY=ds_live_... \
  --env DEEPSIEVE_API_URL=https://deepsieve.ai \
  -- uvx deepsieve-mcp

Manual mcpServers JSON (Cursor, Claude Desktop):

{
  "mcpServers": {
    "deepsieve": {
      "command": "uvx",
      "args": ["deepsieve-mcp"],
      "env": {
        "DEEPSIEVE_API_KEY": "ds_live_...",
        "DEEPSIEVE_API_URL": "https://deepsieve.ai"
      }
    }
  }
}

DEEPSIEVE_API_URL defaults to http://localhost:8200 (local dev stack).

Tools

• Orientation: get_account · list_blueprints · get_blueprint · list_members
• Subscription: cancel_subscription (only when your human asked). There is no resume tool on purpose: resuming re-arms a charge, so it is for an admin on the billing page, or an API key with billing:write over the API or CLI.
• Research: estimate_research · start_research · get_run_status · list_runs · cancel_run · retry_run. estimate_research and start_research require blueprint_id (an id from list_blueprints); the API refuses a real run that does not name one.
• Re-research and monitoring: get_reruns · approve_reruns · set_rerun_auto_approve · list_monitoring · set_monitoring
• Data: query_entities · export_dataset · get_column_changes · get_cell_changes · list_changes · undo_change · list_held_cells · apply_cell_finding · keep_cell_value · release_cell · list_conflicting_cells · choose_cell_value · preview_merge · merge_records · remove_link
• Domain knowledge: list_knowledge · add_knowledge · edit_knowledge · retire_knowledge · restore_knowledge · delete_knowledge · get_run_knowledge · get_knowledge_usage. Writing needs knowledge:write and a member role; it changes what later runs are told, never a stored value.
• Reports: list_reports · get_report_columns · create_report · get_report_data
• Chat: ask_research_agent · get_chat_thread · confirm_chat_action · dismiss_chat_action · revert_chat_action
• Blueprints: propose_blueprint_change · create_blueprint · get_blueprint_draft · answer_blueprint_question · adjust_blueprint · approve_blueprint
• Webhooks: list_webhooks

A new draft stops at ready or at awaiting_input. awaiting_input means it has follow-up questions, and only answer_blueprint_question (or approving as it stands) moves it on.

Full API docs: /developers on your deployment (also /llms.txt).