Refine MCP server

• Addresshttps://getrefine.dev/mcpnot in the official registry: answered when asked

No entry in the official MCP registry names this server. It answered on getrefine.dev when the Atlas asked at the address servers are usually mounted on, so it is the vendor's own. Clients and catalogues that list servers from the registry do not show it.

• Stateanswerschecked 2026-10-28
• Tools24
• Server calls itselfrefine
• Server version4.3.2
• Protocol version2025-06-18

On 2026-10-28 the server answered initialize and tools/list without a sign-in. First answer in 180 ms.

• Whether its tools do the product's job has not been judged yet.
• Checked on 6 of the 7 days it takes to judge reliability.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

• INITE MCP scoreafter a weeka score needs 7 daily checks: 6 so far
• Answeredmeasuring6 of the 7 daily checks it takes to print a share
• Median first answer343 ms
• Space in an agent's context≈ 2,262 tokensthe definitions of all its tools (24), which an agent carries on every turn
• Tool definitionsB · 78 of 100
• Typed results0 of 24 toolsdeclared outputSchema: results come typed, read once

• Parameters without a type: in 1 of 24
• Parameters without a description: in 6 of 24
• Not saying whether it only reads or can destroy: 24 of 24

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

• refine_planning

  Read shared project boards, lanes, Goal cards, routing and durable action results. Every node sees the same project state.

• refine_planning_command

  Create or edit shared boards/lanes/cards, move or release a Goal card, or migrate Todo Lists. Read revisions first and reuse request_id on retry. Returns an action receipt; pending actions execute on the Goal owner.

• refine_planning_action

  Inspect a Project Planning action and its waiting reason, selected node and Skill invocations.

• refine_cancel_planning_action

  Cancel a pending planning action. Completed moves and workflow transitions remain as history.

• refine_list_templates

  List editable system prompt Templates, defaults, revisions and variables. Use refine_request for edits and previews through /templates.

• refine_list_skills

  Read reusable Skill instructions, parameters and configuration revisions.

• refine_skill_triggers

  List the automatic and Custom trigger types available to Skills.

• refine_trigger_skill

  Run one Skill with a Custom trigger on the active node, independently of Goals. Discover required inputs at /skills/{skill_id}/inputs. Reuse request_id to retry transport without duplicating work.

• refine_workflow_control

  Move or explicitly integrate a Goal through shared outcome controls. Read the current workflow_revision first. No Skill is required; forced actions retain override evidence.

• refine_delete_round

  Explicitly delete a Goal Round and all its records, stop current agents, and park the Goal in Backlog. Read the Goal revision first. round_idx is zero-based; use workflow control to resubmit the remaining Round.

• refine_hub_metrics

  Read the saved Metrics Hub snapshot for the target app, including day, week, month, quarter, and year views.

• refine_refresh_hub_metrics

  Calculate and save Metrics Hub aggregates from retained Goal records. Does not launch an agent or change Goals.

• refine_hub_sites

  List Hub sites. Manage assets, collections, records, indexes, queries and publication through /api/hub routes with refine_request.

• refine_system_status

  Read the local Refine daemon status: health, worker state, active operations, and target app state.

• refine_api_groups

  List the daemon API capability groups and what each one covers. Use this to discover routes for the refine_request tool.

• refine_dashboard

  Read the active project dashboard projection and diagnostic summary.

• refine_list_goals

  List Goals (units of product feedback and work) for the active app.

• refine_show_goal

  Show authoritative Goal detail by id, including Round implementation plans, evidence, and current workflow state.

• refine_stop_process

  Stop one managed agent or other process through the shared process-control capability. Linked Goals are cancelled only after exact process exit is confirmed.

• refine_export_goal_jira

  Export a Goal's SOC 2 delivery evidence as Jira-importable CSV, including implementation reports and commits.

• refine_draft_goal

  Draft exactly one implementation-ready Goal from a Plan transcript for review. This does not persist the Goal.

• refine_list_features

  List Features (grouped Goals) for the active app.

• refine_next

  Recommend the next operations for the active project and its fleet, each with the exact CLI command to run. Start here when deciding what to do: it reports failed nodes, work that should be distributed, and reviewable work waiting to…

• refine_request

  Escape hatch: call any local daemon API route directly. Provide 'method' (default GET), 'path' (e.g. /work/goals), and an optional JSON 'body'. Discover routes with the refine_api_groups tool.