Skip to content
OperatorNest

OperatorNest tools over MCP

Connect your assistant to five calculators at https://operatornest.com/mcp. Each tool returns a complete calculation and a link to open the inputs on its matching page.

Server connection

Connect to https://operatornest.com/mcp over Streamable HTTP, without an account, key or sign-in. Calculations use no AI calls and store or log no inputs, though your assistant or gateway may retain them. Client guides below were checked 2 October 2026.

Each request accepts one MCP message up to 32,768 bytes, with a limit of 300 requests per client IP per minute and a 429 response with a 60-second retry interval when you exceed it, so a client can wait before sending the next calculation. Results include units, assumptions and source dates. The link's URL fragment holds inputs for anyone you share it with to read.

The server is open source under the MIT license: OperatorNest/mcp-tools. You can run it yourself over stdio or as a Cloudflare Worker.

Available MCP tools

AI API cost

ai_api_cost: Estimate API workload costs in USD per request, day and month, with rates and dated snapshot provenance. Use for token budgets; use ai_subscription_stack_cost for chat or coding subscriptions. Model IDs use providerId/id from https://operatornest.com/api/models?subset=pricing, filtered to the MCP catalog; the OSS bundle lists accepted IDs in src/data/pricing-snapshot.json. Rates stay offline until the catalog is updated; missing rates return null. Cached tokens are part of total input; published batch and long-context rates apply only when eligible. Non-token charges are excluded.

Open the calculator.

Input schema
{
  "type": "object",
  "additionalProperties": false,
  "properties": {
    "inputTokens": {
      "type": "integer",
      "minimum": 0,
      "maximum": 100000000,
      "description": "Total input tokens per request, including cached input tokens.",
      "examples": [
        10000
      ]
    },
    "cachedInputTokens": {
      "type": "integer",
      "minimum": 0,
      "maximum": 100000000,
      "description": "Cached portion of inputTokens per request; must not exceed inputTokens. Use 0 without caching.",
      "examples": [
        5000
      ]
    },
    "outputTokens": {
      "type": "integer",
      "minimum": 0,
      "maximum": 100000000,
      "description": "Generated output tokens per request.",
      "examples": [
        2000
      ]
    },
    "requestsPerDay": {
      "type": "integer",
      "minimum": 0,
      "maximum": 10000000,
      "description": "Requests used to scale the daily estimate, independent of requestsPerMonth.",
      "examples": [
        100
      ]
    },
    "requestsPerMonth": {
      "type": "integer",
      "minimum": 0,
      "maximum": 100000000,
      "description": "Requests used to scale the monthly estimate; not inferred from the daily count.",
      "examples": [
        3000
      ]
    },
    "modelIds": {
      "type": "array",
      "minItems": 1,
      "maxItems": 20,
      "uniqueItems": true,
      "description": "Unique providerId/id values from the MCP pricing catalog. Unknown or unavailable IDs fail validation.",
      "items": {
        "type": "string",
        "minLength": 1,
        "maxLength": 200,
        "pattern": "\\S",
        "description": "Model ID accepted by the MCP pricing catalog."
      }
    },
    "batchModelIds": {
      "type": "array",
      "minItems": 0,
      "maxItems": 20,
      "uniqueItems": true,
      "items": {
        "type": "string",
        "minLength": 1,
        "maxLength": 200,
        "pattern": "\\S",
        "description": "Selected model ID with a published batch discount."
      },
      "description": "Optional subset of modelIds to price with published batch discounts. Omission or [] uses standard rates; IDs without a published discount are rejected.",
      "examples": [
        []
      ]
    }
  },
  "required": [
    "inputTokens",
    "cachedInputTokens",
    "outputTokens",
    "requestsPerDay",
    "requestsPerMonth",
    "modelIds"
  ]
}

Meeting time zones

meeting_time_zones: Find shared 30-minute meeting slots inside 09:00–17:00 for 2–8 participants. Use to compare workday overlap; use recurring_schedule for repeating task calendars or cron_next_runs for scheduler previews. Returns all candidates, overlap count and up to three best slots ranked nearest to each person’s 13:00, with UTC starts and local labels. The date belongs to the first participant’s zone; daylight saving uses runtime IANA data. No calendar availability or holiday lookup, invitations or calendar writes.

Open the calculator.

Input schema
{
  "type": "object",
  "additionalProperties": false,
  "properties": {
    "date": {
      "type": "string",
      "format": "date",
      "minLength": 10,
      "maxLength": 10,
      "pattern": "^\\d{4}-\\d{2}-\\d{2}$",
      "description": "Meeting date in the first person’s local time zone, from 2000-01-01 to 2100-12-31.",
      "examples": [
        "2026-10-05"
      ]
    },
    "people": {
      "type": "array",
      "minItems": 2,
      "maxItems": 8,
      "description": "Participants in display order. The first person determines the local date; all must fit a full half-hour inside 09:00–17:00.",
      "items": {
        "type": "object",
        "additionalProperties": false,
        "properties": {
          "name": {
            "type": "string",
            "minLength": 1,
            "maxLength": 80,
            "pattern": "\\S",
            "description": "Participant label used in the returned local-time display.",
            "examples": [
              "Sam"
            ]
          },
          "zone": {
            "type": "string",
            "minLength": 1,
            "maxLength": 100,
            "pattern": "\\S",
            "description": "IANA time zone supported by Intl.DateTimeFormat.",
            "examples": [
              "America/New_York",
              "Asia/Kolkata"
            ]
          }
        },
        "required": [
          "name",
          "zone"
        ]
      }
    }
  },
  "required": [
    "date",
    "people"
  ]
}

Recurring schedule

recurring_schedule: Prepare weekly, biweekly or monthly task calendars and estimate monthly effort for up to 20 tasks. Use to draft a recurring work plan; use cron_next_runs to preview existing cron syntax or meeting_time_zones for shared workday slots. Returns monthly hours, first dates, recurrence rules and an RFC 5545 calendar. Times float in the importing calendar’s local zone; weekly estimates use 52/12 runs per month and biweekly 26/12. owner is a planning label, with no assignment or execution. No calendar is updated; an empty task list returns zero hours and no events.

Open the calculator.

Input schema
{
  "type": "object",
  "additionalProperties": false,
  "properties": {
    "startDate": {
      "type": "string",
      "format": "date",
      "minLength": 10,
      "maxLength": 10,
      "pattern": "^\\d{4}-\\d{2}-\\d{2}$",
      "description": "Earliest date for the first occurrence of each task, from 2000-01-01 to 2100-12-31.",
      "examples": [
        "2026-10-05"
      ]
    },
    "tasks": {
      "type": "array",
      "minItems": 0,
      "maxItems": 20,
      "description": "Tasks to include in the estimate and calendar. [] returns zero hours and a calendar with no events.",
      "items": {
        "type": "object",
        "additionalProperties": false,
        "properties": {
          "title": {
            "type": "string",
            "minLength": 1,
            "maxLength": 80,
            "pattern": "\\S",
            "description": "Task label used as the calendar event summary.",
            "examples": [
              "Investor update"
            ]
          },
          "frequency": {
            "type": "string",
            "enum": [
              "weekly",
              "biweekly",
              "monthly"
            ],
            "description": "Repeat every week, every two weeks from the first matching weekday, or every month."
          },
          "day": {
            "type": "string",
            "enum": [
              "MO",
              "TU",
              "WE",
              "TH",
              "FR",
              "SA",
              "SU"
            ],
            "description": "Weekday for weekly/biweekly tasks. Required but ignored for monthly tasks.",
            "examples": [
              "MO"
            ]
          },
          "monthday": {
            "type": "integer",
            "minimum": 1,
            "maximum": 28,
            "description": "Day of month for monthly tasks, limited to dates present in every month. Required but ignored for weekly/biweekly tasks.",
            "examples": [
              1
            ]
          },
          "minutes": {
            "type": "integer",
            "minimum": 1,
            "maximum": 1440,
            "description": "Estimated minutes per run; used for event duration and monthly hours.",
            "examples": [
              35
            ]
          },
          "time": {
            "type": "string",
            "minLength": 5,
            "maxLength": 5,
            "pattern": "^([01][0-9]|2[0-3]):[0-5][0-9]$",
            "description": "24-hour HH:mm start time, floating in the importing calendar’s local time zone.",
            "examples": [
              "09:00"
            ]
          },
          "owner": {
            "type": "string",
            "enum": [
              "delegate",
              "keep",
              "decide"
            ],
            "description": "Planning label: delegate to an operator, keep for yourself, or decide later. Included in the calendar note; does not assign or execute work."
          }
        },
        "required": [
          "title",
          "frequency",
          "day",
          "monthday",
          "minutes",
          "time",
          "owner"
        ]
      }
    }
  },
  "required": [
    "startDate",
    "tasks"
  ]
}

AI subscription stack cost

ai_subscription_stack_cost: Total selected AI chat and coding subscriptions in USD and identify overlapping capabilities. Use to review a subscription stack; use ai_api_cost for API token charges. Select unique planIds from the schema enum; [] returns zero totals and no plans. Returns monthly and annual totals, selected plans, capability groups and dated official sources. Annual totals use published annual billing or twelve monthly charges; taxes, regional differences and promotions are excluded. Bundled rates are dated estimates, without live billing lookup or subscription changes.

Open the calculator.

Input schema
{
  "type": "object",
  "additionalProperties": false,
  "properties": {
    "planIds": {
      "type": "array",
      "minItems": 0,
      "maxItems": 18,
      "uniqueItems": true,
      "description": "Unique subscription IDs from the enum. An empty list returns zero totals and no selected plans.",
      "examples": [
        [
          "chatgpt-plus",
          "claude-pro"
        ],
        []
      ],
      "items": {
        "type": "string",
        "enum": [
          "chatgpt-go",
          "chatgpt-plus",
          "chatgpt-pro",
          "claude-pro",
          "claude-max-5x",
          "claude-max-20x",
          "google-ai-pro",
          "google-ai-ultra-5x",
          "google-ai-ultra-20x",
          "supergrok",
          "supergrok-plus",
          "perplexity-pro",
          "perplexity-max",
          "microsoft-365-premium",
          "mistral-pro",
          "manus-pro",
          "genspark-plus",
          "cursor-pro"
        ],
        "description": "ID of a plan in the bundled subscription catalog."
      }
    }
  },
  "required": [
    "planIds"
  ]
}

Cron next runs

cron_next_runs: Preview up to 10 cron matches strictly after a supplied UTC instant, within a 366-day search. Use to check existing scheduler syntax; use recurring_schedule to draft task recurrence calendars or meeting_time_zones for participant overlap. Returns UTC starts, local labels, normalized inputs and a dated syntax source; fewer matches or [] means the search window ended. Supports numeric stars, lists, ranges, steps and weekday names in the selected dialect; rejects L, W, # and named months. cloudflare requires UTC; other zones follow runtime daylight saving transitions. No schedule is installed or run.

Open the calculator.

Input schema
{
  "type": "object",
  "additionalProperties": false,
  "properties": {
    "expression": {
      "type": "string",
      "minLength": 1,
      "maxLength": 200,
      "pattern": "\\S",
      "description": "Cron fields in the selected dialect: five for github/cloudflare, six or seven for quartz, six for aws (optional cron(...) wrapper). Supports *, lists, ranges, steps and weekday names; rejects L, W, # and named months. Quartz/aws require ? in exactly one day field.",
      "examples": [
        "0 9 * * 1-5"
      ]
    },
    "dialect": {
      "type": "string",
      "enum": [
        "github",
        "cloudflare",
        "quartz",
        "aws"
      ],
      "description": "Scheduler syntax and weekday numbering. github uses 0 or 7 for Sunday and 1–6 for Monday–Saturday; the others use 1–7 for Sunday–Saturday.",
      "examples": [
        "github"
      ]
    },
    "zone": {
      "type": "string",
      "minLength": 1,
      "maxLength": 100,
      "pattern": "\\S",
      "description": "IANA time zone for matching local cron fields. cloudflare requires UTC. Daylight saving transitions follow the runtime time-zone database.",
      "examples": [
        "UTC",
        "America/New_York"
      ]
    },
    "after": {
      "type": "string",
      "format": "date-time",
      "minLength": 20,
      "maxLength": 24,
      "pattern": "^\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}(?:\\.\\d{1,3})?Z$",
      "description": "UTC ISO instant from 2000 through 2100; matches are strictly after this instant. Only Z offsets are accepted.",
      "examples": [
        "2026-10-02T00:00:00Z"
      ]
    },
    "count": {
      "type": "integer",
      "minimum": 1,
      "maximum": 10,
      "description": "Maximum matches to return. The list may be shorter or empty if no further matches occur within the 366-day search.",
      "examples": [
        3
      ]
    }
  },
  "required": [
    "expression",
    "dialect",
    "zone",
    "after",
    "count"
  ]
}

ChatGPT

Enable Developer mode in Settings → Security and login, then add the server URL in Plugins with no authentication and install it from your personal plugins. In a Work chat, type @ and select OperatorNest tools. Check account and workspace access in the official connection guide.

Claude

On Pro or Max, open Customize → Connectors → Add → Add custom connector. Enter “OperatorNest tools” and the server URL, choose No sign in and enable it for the conversation; Team and Enterprise need an authorized owner to add it in Organization settings first, as the official custom connector steps explain.

Codex

Add the server with the command below. Use /mcp to check it; the CLI, desktop app and IDE extension share configuration on the same host. See the official MCP guide.

codex mcp add operatornest-tools --url https://operatornest.com/mcp

Cursor

Merge this entry into .cursor/mcp.json for a project or ~/.cursor/mcp.json for all projects. Save and restart Cursor. See the official manual setup guide.

{
  "mcpServers": {
    "operatornest-tools": {
      "url": "https://operatornest.com/mcp"
    }
  }
}

Gemini CLI

Add the HTTP server below, then check it with gemini mcp list. See the official MCP server guide.

gemini mcp add --transport http operatornest-tools https://operatornest.com/mcp

Other MCP clients

Create a remote HTTP connection named OperatorNest tools with Streamable HTTP, https://operatornest.com/mcp and no authentication. Refresh the list. All five tools should appear. The endpoint supports the current revision and stateless compatibility for older clients under the official MCP transport guide.

Limits and errors

These tools calculate only. They don't read calendars, cancel subscriptions, install schedules or execute tasks. API estimates name the bundled catalog's date. Subscription sources were checked 27 September 2026 and cron sources 28 September 2026. Review those dates before relying on the result.

For invalid inputs, check the schema and retry, using the HTTP status to distinguish a malformed request or batch (400), disallowed browser origin (403), oversized body (413), unsupported content type (415), request limit (429) or unavailable calculation (503) before deciding what the client should change. Check the cause first. Native clients can connect without an Origin header; browser origins are limited to this site, ChatGPT and Claude. You can also use the public JSON API.

Hand off your first task tonight.

Tell us your email and what you'd hand off first. We'll send your access details and help you set up your operator.