plyceMCPMCP verbinden
Recruiting-Kontext für deine KI

Dein CRM.
Im Gespräch
mit deiner KI.

Verbinde plyce mit deinem KI-Assistenten. Stelle Recruiter-Fragen in einem Call – aktive Projekte, Pipeline mit Namen, Interviews und Nachfassen – statt Tabellenbeziehungen zusammenzubauen.

Jetzt verbinden Möglichkeiten entdecken

Model Context Protocol · OAuth · Katalog v6 · Tenant-Isolation

PLYCE / CONTEXT IN ACTION
Beispiel · keine Live-Daten
Welche aktiven Projekte habe ich bei medac?
list_projects
status: Aktivowner_id: mecompany: medac
Ein Call. Fertige Karten.
01ProjektTitel & Status
02KundeCompany-Karte
03Pipelinelist_applications + include
04Nachfassenget_attention_items
Quellen statt Vermutungen.Deine Rechte gelten.

Ein Endpoint. Dein Recruiting-Kontext.Für Remote-MCP-Clients mit OAuth-Unterstützung.

https://mcp.plyce.app/mcp

Weniger Calls.
Mehr Recruiter-Semantik.

Deine KI stellt Recruiting-Fragen. plyce liefert gejointe Karten, Overviews und Nachfassen – strukturiert und nachvollziehbar.

01 / LISTEN

Projekte & Pipeline in einem Call.

Aktive Projekte nach Owner und Kunde. Bewerbungen mit Kandidat, Projekt und Company-Karte – ohne N+1-Lookups.

list_projects · list_applications
02 / STAND

Overview statt Beziehungs-Puzzle.

Projekt-, Kunden- und Kandidaten-Historie mit Pipeline-Counts, Tasks und nächsten Interviews – fest strukturiert.

get_project_overview · get_company_overview
03 / NACHFASSEN

Was heute Aufmerksamkeit braucht.

Deterministische Regeln für überfällige Tasks, stille Sent Outs und ruhende Projekte – mit offengelegter Definition.

get_attention_items · list_scheduled_interviews
So funktioniert MCP

Deine Frage.
Der richtige Kontext.

Das Model Context Protocol ist die gemeinsame Schnittstelle zwischen deinem KI-Client und plyce. Es beschreibt, welche Werkzeuge verfügbar sind und wie sie aufgerufen werden.

Der KI-Client formuliert Antworten und Bewertungen. Der MCP-Server stellt CRM-Daten bereit und prüft bei jedem Abruf deine Zugriffsrechte.

1

Verbinden & anmelden

Dein Client startet die OAuth-Anmeldung. Die Identität wird deinem bestehenden Plyce-Benutzer zugeordnet.

2

Gezielt Werkzeuge aufrufen

Die KI entdeckt erlaubte Tools, sucht Datensätze und folgt den freigegebenen Beziehungen.

3

Mit belegten Daten antworten

Ergebnisse enthalten IDs, Kontext und, wo verfügbar, einen Plyce-Link. Die KI kann Informationen nachladen und Quellen nennen.

In deinem Client.
Mit deinem Plyce-Zugang.

Du benötigst einen aktiven Plyce-Benutzer und einen Client, der Remote MCP mit OAuth unterstützt.

Über die MCP-Einstellungen

  1. Öffne in deinem KI-Client die Einstellungen für Apps, Integrationen oder MCP-Server. Die Bezeichnung hängt vom Client ab.
  2. Füge plyce als Remote-Server mit dieser URL hinzu:
    https://mcp.plyce.app/mcp
  3. Starte die angebotene OAuth-Anmeldung und melde dich mit deiner verknüpften Identität an.
  4. Aktiviere plyce im gewünschten Chat oder starte eine neue Sitzung.
Eine Verbindung in Codex wird nicht automatisch zu einer App-Verbindung in ChatGPT oder einem anderen Client.

Mit der Codex CLI

Wenn die CLI bereits installiert ist:

codex mcp add plyce --url https://mcp.plyce.app/mcp
codex mcp login plyce
codex mcp list

Unter Windows PowerShell kannst du codex.cmd statt codex verwenden. Nach erfolgreicher Anmeldung erscheint in der Liste OAuth.

Zum Testen: „Welche aktiven Projekte habe ich? Zeige Interviews nächste Woche und wo ich nachfassen muss.“
Kontrolle bleibt bei dir

Mehr Kontext.
Klare Grenzen.

Ein Tenant pro Kontext

Der Server bestimmt den Tenant aus deiner Identität. Datensätze und verknüpfte Ziele werden darauf eingeschränkt.

Deine bestehenden Rechte

OAuth-Scopes, Plyce-Berechtigungen und Datenbankregeln bestimmen gemeinsam, welche Inhalte du lesen kannst.

Inhalte gezielt freigeben

Volltexte, Dokumente und strukturierte Finanzfelder haben eigene MCP-Berechtigungen. Technische Zugangstokens werden nicht ausgegeben.

Aktionen bewusst aktivieren

Schreibwerkzeuge nur mit Tenant-Freigabe (kein separates E-Mail-Flag). Outbound-Mail action-basiert — heute Absage über send_rejection; weitere Send-Actions können folgen. prepare_email bleibt Entwurf. Kontakte brauchen Unternehmen (ID oder eindeutige Domain). Quota: 200/Stunde/User über alle Tools.

Der Tool-Katalog.
Direkt aus dem Server.

Hier siehst du den implementierten Katalog. Deine tatsächlich erlaubten Werkzeuge liefert nach der Anmeldung tools/list.

61 Werkzeuge
list_usersLesen

Search tenant users by name, returning user_id, profile_id, name and active status. Paginated. Prefer resolve_user for single-person lookup.

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {
    "query": {
      "type": "string"
    },
    "include_inactive": {
      "type": "boolean"
    },
    "limit": {
      "type": "integer",
      "description": "1..100"
    },
    "cursor": {
      "type": "string",
      "description": "Opaque next_cursor, bound to actor and all filters."
    }
  },
  "required": [],
  "additionalProperties": false
}
resolve_userLesen

Resolve a name, user_id or profile_id. Returns status resolved|ambiguous|not_found with matches — never auto-select an ambiguous name. Use me on other tools for the signed-in user.

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {
    "name": {
      "type": "string"
    },
    "user_id": {
      "type": "string"
    },
    "profile_id": {
      "type": "string"
    }
  },
  "required": [],
  "additionalProperties": false
}
get_user_activity_reportLesen

Alias of get_activity_report with group_by=user. Prefer get_activity_report(group_by=user, include_rows=...).

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp, reporting:read
Plyce-Rechte
reporting.view
{
  "type": "object",
  "properties": {
    "period": {
      "type": "string",
      "enum": [
        "today",
        "yesterday",
        "this_week",
        "last_week",
        "last_7_days",
        "this_month",
        "last_month",
        "this_quarter",
        "last_quarter",
        "this_year"
      ],
      "description": "Use period OR from + to. Current periods end after today, Europe/Berlin."
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or ISO timestamp with explicit UTC offset."
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or exclusive ISO timestamp with explicit offset. Same format as from."
    },
    "user_ids": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Auth user UUIDs; resolve names/profile IDs through resolve_user first."
    },
    "event_keys": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "exclude_automated": {
      "type": "boolean"
    },
    "exclude_system": {
      "type": "boolean"
    },
    "origins": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "manual",
          "automated",
          "system",
          "bot",
          "bulk",
          "unknown"
        ]
      }
    },
    "include_activities": {
      "type": "boolean"
    },
    "limit": {
      "type": "integer",
      "description": "1..100"
    },
    "cursor": {
      "type": "string",
      "description": "Opaque next_cursor, bound to actor and all filters."
    }
  },
  "required": [],
  "additionalProperties": false
}
get_activity_reportLesen

Activity KPIs with optional group_by (user|event_key|day|week|month), interval buckets and include_rows for drill-down. Replaces get_user_activity_report, list_report_activities and get_time_report.

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp, reporting:read
Plyce-Rechte
reporting.view
{
  "type": "object",
  "properties": {
    "period": {
      "type": "string",
      "enum": [
        "today",
        "yesterday",
        "this_week",
        "last_week",
        "last_7_days",
        "this_month",
        "last_month",
        "this_quarter",
        "last_quarter",
        "this_year"
      ],
      "description": "Use period OR from + to. Current periods end after today, Europe/Berlin."
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or ISO timestamp with explicit UTC offset."
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or exclusive ISO timestamp with explicit offset. Same format as from."
    },
    "user_ids": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Auth user UUIDs; resolve names/profile IDs through resolve_user first."
    },
    "event_keys": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "exclude_automated": {
      "type": "boolean"
    },
    "exclude_system": {
      "type": "boolean"
    },
    "origins": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "manual",
          "automated",
          "system",
          "bot",
          "bulk",
          "unknown"
        ]
      }
    },
    "group_by": {
      "type": "string",
      "enum": [
        "user",
        "event_key",
        "day",
        "week",
        "month"
      ]
    },
    "interval": {
      "type": "string",
      "enum": [
        "day",
        "week",
        "month"
      ]
    },
    "include_rows": {
      "type": "boolean"
    },
    "limit": {
      "type": "integer",
      "description": "1..100"
    },
    "cursor": {
      "type": "string",
      "description": "Opaque next_cursor, bound to actor and all filters."
    }
  },
  "required": [],
  "additionalProperties": false
}
list_report_activitiesLesen

Alias of get_activity_report(include_rows=true). Prefer the unified report tool.

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp, reporting:read
Plyce-Rechte
reporting.view, activities.view
{
  "type": "object",
  "properties": {
    "period": {
      "type": "string",
      "enum": [
        "today",
        "yesterday",
        "this_week",
        "last_week",
        "last_7_days",
        "this_month",
        "last_month",
        "this_quarter",
        "last_quarter",
        "this_year"
      ],
      "description": "Use period OR from + to. Current periods end after today, Europe/Berlin."
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or ISO timestamp with explicit UTC offset."
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or exclusive ISO timestamp with explicit offset. Same format as from."
    },
    "user_ids": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Auth user UUIDs; resolve names/profile IDs through resolve_user first."
    },
    "event_keys": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "exclude_automated": {
      "type": "boolean"
    },
    "exclude_system": {
      "type": "boolean"
    },
    "origins": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "manual",
          "automated",
          "system",
          "bot",
          "bulk",
          "unknown"
        ]
      }
    },
    "limit": {
      "type": "integer",
      "description": "1..100"
    },
    "cursor": {
      "type": "string",
      "description": "Opaque next_cursor, bound to actor and all filters."
    }
  },
  "required": [],
  "additionalProperties": false
}
get_time_reportLesen

Alias of get_activity_report(interval=day|week|month). Prefer the unified report tool.

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp, reporting:read
Plyce-Rechte
reporting.view
{
  "type": "object",
  "properties": {
    "period": {
      "type": "string",
      "enum": [
        "today",
        "yesterday",
        "this_week",
        "last_week",
        "last_7_days",
        "this_month",
        "last_month",
        "this_quarter",
        "last_quarter",
        "this_year"
      ],
      "description": "Use period OR from + to. Current periods end after today, Europe/Berlin."
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or ISO timestamp with explicit UTC offset."
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or exclusive ISO timestamp with explicit offset. Same format as from."
    },
    "user_ids": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Auth user UUIDs; resolve names/profile IDs through resolve_user first."
    },
    "event_keys": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "exclude_automated": {
      "type": "boolean"
    },
    "exclude_system": {
      "type": "boolean"
    },
    "origins": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "manual",
          "automated",
          "system",
          "bot",
          "bulk",
          "unknown"
        ]
      }
    },
    "bucket": {
      "type": "string",
      "enum": [
        "day",
        "week",
        "month"
      ]
    }
  },
  "required": [],
  "additionalProperties": false
}
get_pipeline_reportLesen

mode=snapshot (default without dates): current active-project pipeline counts. mode=events: historical status-change events (requires period or from/to). Replaces list_pipeline_transitions.

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp, reporting:read
Plyce-Rechte
reporting.view
{
  "type": "object",
  "properties": {
    "period": {
      "type": "string",
      "enum": [
        "today",
        "yesterday",
        "this_week",
        "last_week",
        "last_7_days",
        "this_month",
        "last_month",
        "this_quarter",
        "last_quarter",
        "this_year"
      ],
      "description": "Use period OR from + to. Current periods end after today, Europe/Berlin."
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or ISO timestamp with explicit UTC offset."
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or exclusive ISO timestamp with explicit offset. Same format as from."
    },
    "user_ids": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Auth user UUIDs; resolve names/profile IDs through resolve_user first."
    },
    "event_keys": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "exclude_automated": {
      "type": "boolean"
    },
    "exclude_system": {
      "type": "boolean"
    },
    "origins": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "manual",
          "automated",
          "system",
          "bot",
          "bulk",
          "unknown"
        ]
      }
    },
    "project_ids": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "company_ids": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "status": {
      "type": "string"
    },
    "mode": {
      "type": "string",
      "enum": [
        "snapshot",
        "events"
      ],
      "description": "snapshot = current pipeline counts; events = historical transitions (requires period or from/to)"
    },
    "limit": {
      "type": "integer",
      "description": "1..100"
    },
    "cursor": {
      "type": "string",
      "description": "Opaque next_cursor, bound to actor and all filters."
    }
  },
  "required": [],
  "additionalProperties": false
}
list_pipeline_transitionsLesen

Alias of get_pipeline_report(mode=events). Prefer the unified pipeline tool.

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp, reporting:read
Plyce-Rechte
reporting.view, activities.view
{
  "type": "object",
  "properties": {
    "period": {
      "type": "string",
      "enum": [
        "today",
        "yesterday",
        "this_week",
        "last_week",
        "last_7_days",
        "this_month",
        "last_month",
        "this_quarter",
        "last_quarter",
        "this_year"
      ],
      "description": "Use period OR from + to. Current periods end after today, Europe/Berlin."
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or ISO timestamp with explicit UTC offset."
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or exclusive ISO timestamp with explicit offset. Same format as from."
    },
    "user_ids": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Auth user UUIDs; resolve names/profile IDs through resolve_user first."
    },
    "event_keys": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "exclude_automated": {
      "type": "boolean"
    },
    "exclude_system": {
      "type": "boolean"
    },
    "origins": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "manual",
          "automated",
          "system",
          "bot",
          "bulk",
          "unknown"
        ]
      }
    },
    "project_ids": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "company_ids": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "status": {
      "type": "string"
    },
    "limit": {
      "type": "integer",
      "description": "1..100"
    },
    "cursor": {
      "type": "string",
      "description": "Opaque next_cursor, bound to actor and all filters."
    }
  },
  "required": [],
  "additionalProperties": false
}
get_placementsLesen

Historical Placement events and authorized relations. Financial details only with mcp.financial.read.

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp, reporting:read
Plyce-Rechte
reporting.view, activities.view
{
  "type": "object",
  "properties": {
    "period": {
      "type": "string",
      "enum": [
        "today",
        "yesterday",
        "this_week",
        "last_week",
        "last_7_days",
        "this_month",
        "last_month",
        "this_quarter",
        "last_quarter",
        "this_year"
      ],
      "description": "Use period OR from + to. Current periods end after today, Europe/Berlin."
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or ISO timestamp with explicit UTC offset."
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or exclusive ISO timestamp with explicit offset. Same format as from."
    },
    "user_ids": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Auth user UUIDs; resolve names/profile IDs through resolve_user first."
    },
    "event_keys": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "exclude_automated": {
      "type": "boolean"
    },
    "exclude_system": {
      "type": "boolean"
    },
    "origins": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "manual",
          "automated",
          "system",
          "bot",
          "bulk",
          "unknown"
        ]
      }
    },
    "project_ids": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "company_ids": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "limit": {
      "type": "integer",
      "description": "1..100"
    },
    "cursor": {
      "type": "string",
      "description": "Opaque next_cursor, bound to actor and all filters."
    }
  },
  "required": [],
  "additionalProperties": false
}
get_deal_value_reportLesen

Placement deal values with recruiter splits, consistent with Plyce deal-value reporting. NOT invoiced revenue, cash receipts or realized margin.

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp, reporting:read
Plyce-Rechte
reporting.view, activities.view, mcp.financial.read
{
  "type": "object",
  "properties": {
    "period": {
      "type": "string",
      "enum": [
        "today",
        "yesterday",
        "this_week",
        "last_week",
        "last_7_days",
        "this_month",
        "last_month",
        "this_quarter",
        "last_quarter",
        "this_year"
      ],
      "description": "Use period OR from + to. Current periods end after today, Europe/Berlin."
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or ISO timestamp with explicit UTC offset."
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or exclusive ISO timestamp with explicit offset. Same format as from."
    },
    "user_ids": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Auth user UUIDs; resolve names/profile IDs through resolve_user first."
    },
    "event_keys": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "exclude_automated": {
      "type": "boolean"
    },
    "exclude_system": {
      "type": "boolean"
    },
    "origins": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "manual",
          "automated",
          "system",
          "bot",
          "bulk",
          "unknown"
        ]
      }
    },
    "project_ids": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "company_ids": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "breakdown": {
      "type": "string",
      "enum": [
        "user",
        "company",
        "project",
        "placement",
        "employment_type"
      ]
    }
  },
  "required": [],
  "additionalProperties": false
}
find_similar_candidatesLesen

Deterministic similarity ranking from structured skills, job title and seniority. Scores disclose weights, evidence, unknown dimensions and search coverage. Uses a bounded literal skill search (no synonym expansion). No LLM.

No cross-tenant access, invented historical facts, AI generation or write actions.

OAuth-Scopes
mcp, candidates:read
Plyce-Rechte
candidates.view
{
  "type": "object",
  "properties": {
    "candidate_id": {
      "type": "string"
    },
    "location": {
      "type": "string"
    },
    "postal_code": {
      "type": "string"
    },
    "radius_km": {
      "type": "number"
    },
    "latitude": {
      "type": "number"
    },
    "longitude": {
      "type": "number"
    },
    "candidate_status": {
      "type": "string"
    },
    "limit": {
      "type": "integer"
    }
  },
  "required": [
    "candidate_id"
  ],
  "additionalProperties": false
}
get_data_catalogLesen

Fallback schema discovery when no business list/overview fits. Prefer list_projects, list_applications and overviews for recruiter questions.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {},
  "required": [],
  "additionalProperties": false
}
get_entityLesen

Read a CRM record by type and id. detail_level=card returns labels and links; full (default) returns allowlisted fields. Prefer this over get_candidate/get_project/get_company/get_contact.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact",
        "application",
        "assignment",
        "job_posting",
        "availability_change",
        "saved_list",
        "staff"
      ]
    },
    "id": {
      "type": "string",
      "description": "Entity UUID"
    },
    "detail_level": {
      "type": "string",
      "enum": [
        "card",
        "full"
      ],
      "description": "card = labels and links; full = allowlisted fields (default full)"
    }
  },
  "required": [
    "type",
    "id"
  ],
  "additionalProperties": false
}
get_applicationLesen

Read an application after checking both candidate and project access. Prefer get_entity(type=application).

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
candidates.view, projects.view
{
  "type": "object",
  "properties": {
    "id": {
      "type": "string"
    }
  },
  "required": [
    "id"
  ],
  "additionalProperties": false
}
list_projectsLesen

Page projects with filters for status, owner (me), sourcing user, company, employment type, work mode and updated_since. Returns project cards with company when accessible.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
projects.view
{
  "type": "object",
  "properties": {
    "limit": {
      "type": "integer",
      "description": "1..100, default 25"
    },
    "cursor": {
      "type": "string",
      "description": "next_cursor from the same tool, filters and entity"
    },
    "status": {
      "type": "string"
    },
    "owner_id": {
      "type": "string",
      "description": "Project owner user UUID, or me for the signed-in user"
    },
    "sourcing_user_id": {
      "type": "string",
      "description": "User UUID in sourcing_users, or me"
    },
    "company_id": {
      "type": "string"
    },
    "anstellungsart": {
      "type": "string"
    },
    "work_mode": {
      "type": "string"
    },
    "updated_since": {
      "type": "string",
      "description": "ISO date/time; inclusive lower bound on updated_at"
    }
  },
  "required": [],
  "additionalProperties": false
}
list_applicationsLesen

Page applications with recruiter filters (company, project status/owner, recruiter, candidate status) and joined candidate/project/company cards so Interview lists need no N+1 lookups. For “currently interviewing” prefer project_status=Aktiv.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
candidates.view, projects.view
{
  "type": "object",
  "properties": {
    "limit": {
      "type": "integer",
      "description": "1..100, default 25"
    },
    "cursor": {
      "type": "string",
      "description": "next_cursor from the same tool, filters and entity"
    },
    "candidate_id": {
      "type": "string"
    },
    "project_id": {
      "type": "string"
    },
    "company_id": {
      "type": "string"
    },
    "status": {
      "type": "string"
    },
    "project_status": {
      "type": "string"
    },
    "recruiter_id": {
      "type": "string",
      "description": "Candidate recruiter user UUID, or me"
    },
    "project_owner_id": {
      "type": "string",
      "description": "Project owner user UUID, or me"
    },
    "candidate_status": {
      "type": "string"
    },
    "updated_since": {
      "type": "string",
      "description": "ISO date/time; inclusive lower bound on application updated_at"
    },
    "include": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Join cards: candidate, project, company. Default all three."
    }
  },
  "required": [],
  "additionalProperties": false
}
list_scheduled_interviewsLesen

List calendar interviews from activity interview.date/time metadata — not pipeline status Interview. Returns date, time, round, candidate, project and company. One call fills a page or exhausts matches (no empty has_more pages).

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
candidates.view, projects.view, activities.view
{
  "type": "object",
  "properties": {
    "limit": {
      "type": "integer",
      "description": "1..100, default 25"
    },
    "cursor": {
      "type": "string",
      "description": "next_cursor from the same tool, filters and entity"
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day or ISO timestamp"
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day or exclusive ISO timestamp"
    },
    "recruiter_id": {
      "type": "string",
      "description": "Actor user UUID on the interview activity, or me"
    },
    "project_id": {
      "type": "string"
    },
    "company_id": {
      "type": "string"
    },
    "candidate_id": {
      "type": "string"
    }
  },
  "required": [],
  "additionalProperties": false
}
get_project_overviewLesen

One-call project stand: project, company, hiring manager, owner, sourcing team, pipeline counts, candidates by status, recent activities, open tasks, next interviews.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
projects.view
{
  "type": "object",
  "properties": {
    "project_id": {
      "type": "string"
    }
  },
  "required": [
    "project_id"
  ],
  "additionalProperties": false
}
get_company_overviewLesen

One-call customer stand: company, owner, key contacts, active/on-hold projects, recent placements, pipeline, activities, open tasks.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
companies.view
{
  "type": "object",
  "properties": {
    "company_id": {
      "type": "string"
    }
  },
  "required": [
    "company_id"
  ],
  "additionalProperties": false
}
get_candidate_historyLesen

Candidate history across applications, projects, sent outs, interviews, offers, placements, assignments and recent communication.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
candidates.view
{
  "type": "object",
  "properties": {
    "candidate_id": {
      "type": "string"
    }
  },
  "required": [
    "candidate_id"
  ],
  "additionalProperties": false
}
get_attention_itemsLesen

Deterministic follow-up queue. Default types are lean (no candidate_without_followup / project_without_candidates). Definitions returned with items; cap per type.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {
    "user_id": {
      "type": "string",
      "description": "User UUID or me; defaults to signed-in user"
    },
    "types": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Optional subset. Default (lean): overdue_task, sentout_without_feedback, interview_without_followup, offer_without_update, active_project_without_activity. Opt-in heavy: candidate_without_followup, project_without_candidates."
    },
    "older_than_days": {
      "type": "integer",
      "description": "Days without update/activity; default 7, max 90"
    }
  },
  "required": [],
  "additionalProperties": false
}
list_related_entitiesLesen

Fallback edge traversal from get_data_catalog when no overview or list tool fits. Prefer list_projects, list_applications and overviews first.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact",
        "application",
        "assignment",
        "job_posting",
        "availability_change",
        "saved_list",
        "staff"
      ]
    },
    "id": {
      "type": "string",
      "description": "Entity UUID"
    },
    "limit": {
      "type": "integer",
      "description": "1..100, default 25"
    },
    "cursor": {
      "type": "string",
      "description": "next_cursor from the same tool, filters and entity"
    },
    "relation": {
      "type": "string",
      "description": "Relation name from get_data_catalog"
    }
  },
  "required": [
    "type",
    "id",
    "relation"
  ],
  "additionalProperties": false
}
list_entity_activitiesLesen

Page the activity timeline for one entity (ordered by relation ID). Prefer list_activities. For KPI drill-down use get_activity_report(include_rows=true).

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
activities.view
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact",
        "application",
        "assignment",
        "job_posting",
        "availability_change",
        "saved_list",
        "staff"
      ]
    },
    "id": {
      "type": "string",
      "description": "Entity UUID"
    },
    "limit": {
      "type": "integer",
      "description": "1..100, default 25"
    },
    "cursor": {
      "type": "string",
      "description": "next_cursor from the same tool, filters and entity"
    },
    "period": {
      "type": "string",
      "enum": [
        "today",
        "yesterday",
        "this_week",
        "last_week",
        "last_7_days",
        "this_month",
        "last_month",
        "this_quarter",
        "last_quarter",
        "this_year"
      ],
      "description": "Use period OR from + to. Current periods end after today, Europe/Berlin."
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or ISO timestamp with explicit UTC offset."
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or exclusive ISO timestamp with explicit offset. Same format as from."
    }
  },
  "required": [
    "type",
    "id"
  ],
  "additionalProperties": false
}
list_activitiesLesen

Page the activity timeline for one entity. Alias of list_entity_activities.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
activities.view
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact",
        "application",
        "assignment",
        "job_posting",
        "availability_change",
        "saved_list",
        "staff"
      ]
    },
    "id": {
      "type": "string",
      "description": "Entity UUID"
    },
    "limit": {
      "type": "integer",
      "description": "1..100, default 25"
    },
    "cursor": {
      "type": "string",
      "description": "next_cursor from the same tool, filters and entity"
    },
    "period": {
      "type": "string",
      "enum": [
        "today",
        "yesterday",
        "this_week",
        "last_week",
        "last_7_days",
        "this_month",
        "last_month",
        "this_quarter",
        "last_quarter",
        "this_year"
      ],
      "description": "Use period OR from + to. Current periods end after today, Europe/Berlin."
    },
    "from": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or ISO timestamp with explicit UTC offset."
    },
    "to": {
      "type": "string",
      "description": "YYYY-MM-DD inclusive Berlin day, or exclusive ISO timestamp with explicit offset. Same format as from."
    }
  },
  "required": [
    "type",
    "id"
  ],
  "additionalProperties": false
}
get_activityLesen

Read an activity in an authorized entity context. All targets of shared notes must be readable.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
activities.view
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact",
        "application",
        "assignment",
        "job_posting",
        "availability_change",
        "saved_list",
        "staff"
      ]
    },
    "id": {
      "type": "string",
      "description": "Entity UUID"
    },
    "activity_id": {
      "type": "string"
    }
  },
  "required": [
    "type",
    "id",
    "activity_id"
  ],
  "additionalProperties": false
}
list_entity_documentsLesen

Page document metadata for a candidate, project or company.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
mcp.documents.read
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact",
        "application",
        "assignment",
        "job_posting",
        "availability_change",
        "saved_list",
        "staff"
      ]
    },
    "id": {
      "type": "string",
      "description": "Entity UUID"
    },
    "limit": {
      "type": "integer",
      "description": "1..100, default 25"
    },
    "cursor": {
      "type": "string",
      "description": "next_cursor from the same tool, filters and entity"
    }
  },
  "required": [
    "type",
    "id"
  ],
  "additionalProperties": false
}
read_documentLesen

Read PDF, DOCX or UTF-8 text from authorized CRM storage in versioned chunks. PDF images require OCR outside this tool. For the original file use get_document_download_url.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
mcp.documents.read, mcp.content.read
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact",
        "application",
        "assignment",
        "job_posting",
        "availability_change",
        "saved_list",
        "staff"
      ]
    },
    "id": {
      "type": "string",
      "description": "Entity UUID"
    },
    "document_id": {
      "type": "string"
    },
    "chunk": {
      "type": "integer",
      "description": "Zero-based text chunk; default 0"
    },
    "version": {
      "type": "string",
      "description": "Required after chunk 0; exact content hash from first response"
    }
  },
  "required": [
    "type",
    "id",
    "document_id"
  ],
  "additionalProperties": false
}
get_document_download_urlLesen

Return a short-lived signed URL for the original file (candidate, project or company). Metadata-only list_entity_documents never includes URLs; text extraction stays on read_document.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
mcp.documents.read, mcp.content.read
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact",
        "application",
        "assignment",
        "job_posting",
        "availability_change",
        "saved_list",
        "staff"
      ]
    },
    "id": {
      "type": "string",
      "description": "Entity UUID"
    },
    "document_id": {
      "type": "string"
    }
  },
  "required": [
    "type",
    "id",
    "document_id"
  ],
  "additionalProperties": false
}
list_tasksLesen

Page manual tasks by entity, assignee (me), status, due_before or overdue. Replaces list_open_tasks and get_daily_briefing.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
mail_automation.view
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact",
        "application",
        "assignment",
        "job_posting",
        "availability_change",
        "saved_list",
        "staff"
      ]
    },
    "id": {
      "type": "string",
      "description": "Entity UUID"
    },
    "limit": {
      "type": "integer",
      "description": "1..100, default 25"
    },
    "cursor": {
      "type": "string",
      "description": "next_cursor from the same tool, filters and entity"
    },
    "assigned_to": {
      "type": "string",
      "description": "Assignee user UUID, or me. Defaults to me when overdue is true and assigned_to omitted."
    },
    "status": {
      "type": "string",
      "enum": [
        "open",
        "completed"
      ]
    },
    "due_before": {
      "type": "string",
      "description": "YYYY-MM-DD, inclusive"
    },
    "overdue": {
      "type": "boolean",
      "description": "Open tasks with due_date before today Europe/Berlin"
    }
  },
  "required": [],
  "additionalProperties": false
}
list_saved_listsLesen

Read only this user’s saved lists and permitted filter definitions.

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {
    "limit": {
      "type": "integer",
      "description": "1..100, default 25"
    },
    "cursor": {
      "type": "string",
      "description": "next_cursor from the same tool, filters and entity"
    }
  },
  "required": [],
  "additionalProperties": false
}
search_candidatesLesen

Structured recruiting search: skills, titles, tags, status and radius. For plain name/text lookup use search (search_entities).

No arbitrary SQL, tenant selection, write operation or hidden background AI.

OAuth-Scopes
mcp
Plyce-Rechte
candidates.view
{
  "type": "object",
  "properties": {
    "limit": {
      "type": "integer",
      "description": "1..100, default 25"
    },
    "cursor": {
      "type": "string",
      "description": "next_cursor from the same tool, filters and entity"
    },
    "use_synonyms": {
      "type": "boolean",
      "description": "Default true: disclosed curated related skill terms, not semantic embeddings. Set false for literal skill substrings."
    },
    "skills": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "skills_mode": {
      "type": "string",
      "enum": [
        "any",
        "all"
      ]
    },
    "job_titles": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "keywords": {
      "type": "string"
    },
    "location": {
      "type": "string",
      "description": "Without radius_km: substring match on city or leave empty and use postal_code. With radius_km: used with postal_code to form the search center (München/Berlin/Hamburg centers are built in; other places are geocoded; ambiguous places require explicit coordinates)."
    },
    "postal_code": {
      "type": "string",
      "description": "Without radius_km: substring match on PLZ. With radius_km: combined with location for the distance center."
    },
    "radius_km": {
      "type": "number",
      "description": "Straight-line distance in km on coordinates (>0..1000). Without radius, location/postal_code are text filters only."
    },
    "latitude": {
      "type": "number"
    },
    "longitude": {
      "type": "number"
    },
    "candidate_status": {
      "type": "string"
    },
    "tags": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "career_level": {
      "type": "string",
      "description": "Match candidates that have this career-level tag (e.g. Senior Level, Mid Level). Stored in tags, not a separate column."
    },
    "recruiter_id": {
      "type": "string",
      "description": "Candidate recruiter user UUID, or me"
    },
    "updated_since": {
      "type": "string",
      "description": "ISO date/time"
    },
    "exclude_candidate_ids": {
      "type": "array",
      "items": {
        "type": "string"
      }
    }
  },
  "required": [],
  "additionalProperties": false
}
get_current_user_contextLesen

Return the logged-in plyce user, tenant, scopes, and which tools they may call.

Not a way to impersonate another user or list other tenants.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
get_allowed_valuesLesen

Enums the agent must use (candidate status, pipeline status, activity types).

Not a search and not a schema dump of the database.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
get_mcp_connection_statusLesen

Show this MCP connection (client, scopes, last seen). Admin/diagnostics — not for recruiting questions.

Not an admin dump of every connection in the tenant.

OAuth-Scopes
mcp:manage
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
get_mcp_usageLesen

Show remaining shared hourly quota (200 calls across all tools) for this user and MCP client.

Not billing and not other users’ usage.

OAuth-Scopes
mcp:manage
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
revoke_mcp_connectionLesen

Revoke this MCP connection immediately.

Does not delete CRM data.

OAuth-Scopes
mcp:manage
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
search_entitiesLesen

Name/text search over candidates, projects, companies, contacts. Prefer this over search_candidates for plain name lookup.

Does not rank against a job ad, does not call plyce matching, does not read CVs. The user model ranks locally.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {
    "query": {
      "type": "string",
      "description": "Full-text query (min 2 characters)"
    },
    "types": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "candidate",
          "project",
          "company",
          "contact"
        ]
      },
      "description": "Optional entity filter"
    },
    "cursor": {
      "type": "string",
      "description": "Opaque pagination cursor"
    }
  },
  "required": [
    "query"
  ],
  "additionalProperties": false
}
searchLesen

Alias of search_entities (name/text search). Prefer search_entities; keep search for ChatGPT host compatibility.

Does not rank against a job ad, does not call plyce matching, does not read CVs. The user model ranks locally.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {
    "query": {
      "type": "string",
      "description": "Full-text query (min 2 characters)"
    },
    "types": {
      "type": "array",
      "items": {
        "type": "string",
        "enum": [
          "candidate",
          "project",
          "company",
          "contact"
        ]
      },
      "description": "Optional entity filter"
    },
    "cursor": {
      "type": "string",
      "description": "Opaque pagination cursor"
    }
  },
  "required": [
    "query"
  ],
  "additionalProperties": false
}
fetchLesen

Alias of get_entity via composite id from search. Prefer get_entity(type, id).

Not a SQL proxy and not a document/CV download.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "Composite id like candidate:<uuid>"
    }
  },
  "required": [
    "id"
  ],
  "additionalProperties": false
}
get_candidateLesen

Alias of get_entity(type=candidate, detail_level=card). Prefer get_entity.

No salary, gender, birth date, CV text, or dossier.

OAuth-Scopes
candidates:read
Plyce-Rechte
candidates.view
{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "Candidate UUID"
    }
  },
  "required": [
    "id"
  ],
  "additionalProperties": false
}
get_projectLesen

Alias of get_entity(type=project, detail_level=card). Prefer get_entity.

No salary band or full job-ad HTML (user model must not be fed raw ads as a matching engine).

OAuth-Scopes
projects:read
Plyce-Rechte
projects.view
{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "Project UUID"
    }
  },
  "required": [
    "id"
  ],
  "additionalProperties": false
}
get_companyLesen

Alias of get_entity(type=company, detail_level=card). Prefer get_entity.

Not news scraping or plyce-AI company research.

OAuth-Scopes
companies:read
Plyce-Rechte
companies.view
{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "Company UUID"
    }
  },
  "required": [
    "id"
  ],
  "additionalProperties": false
}
get_contactLesen

Alias of get_entity(type=contact, detail_level=card). Prefer get_entity.

Not an email send.

OAuth-Scopes
contacts:read
Plyce-Rechte
contacts.view
{
  "type": "object",
  "properties": {
    "id": {
      "type": "string",
      "description": "Contact UUID"
    }
  },
  "required": [
    "id"
  ],
  "additionalProperties": false
}
list_open_tasksLesen

Alias of list_tasks(status=open). Prefer list_tasks.

Does not complete or send mail.

OAuth-Scopes
mcp
Plyce-Rechte
mail_automation.view
{
  "type": "object",
  "properties": {
    "assigned_to": {
      "type": "string"
    },
    "limit": {
      "type": "number"
    }
  },
  "additionalProperties": false
}
get_daily_briefingLesen

Alias of list_tasks(overdue=true) for the signed-in user. Prefer list_tasks.

Not an LLM summary and not matching.

OAuth-Scopes
mcp
Plyce-Rechte
mail_automation.view
{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
get_entity_360Lesen

Alias of get_project_overview / get_company_overview / get_candidate_history for those types; otherwise a generic relation pack.

Not a CV or document viewer.

OAuth-Scopes
mcp
Plyce-Rechte
Prüfung je Datensatz und Beziehung
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact"
      ]
    },
    "id": {
      "type": "string"
    },
    "include": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Up to six relation names from get_data_catalog"
    }
  },
  "required": [
    "type",
    "id"
  ],
  "additionalProperties": false
}
add_noteSchreiben

Store a note written by the user model as candidate|contact|company|project.manual_note.

Does not generate the note text. plyce only stores it. One canonical activity row.

OAuth-Scopes
activities:write
Plyce-Rechte
activities.view
{
  "type": "object",
  "properties": {
    "entity_type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact"
      ]
    },
    "entity_id": {
      "type": "string"
    },
    "text": {
      "type": "string",
      "description": "Note body written by the user model, stored as-is"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "entity_type",
    "entity_id",
    "text",
    "idempotency_key"
  ],
  "additionalProperties": false
}
create_taskSchreiben

Create a manual task (same table as the plyce task dialog). Idempotent.

Does not email the assignee and does not log an activity until the task is completed in plyce.

OAuth-Scopes
tasks:write
Plyce-Rechte
mail_automation.view
{
  "type": "object",
  "properties": {
    "entity_type": {
      "type": "string",
      "enum": [
        "candidate",
        "project",
        "company",
        "contact"
      ]
    },
    "entity_id": {
      "type": "string"
    },
    "entity_name": {
      "type": "string"
    },
    "subject": {
      "type": "string"
    },
    "body": {
      "type": "string"
    },
    "due_date": {
      "type": "string",
      "description": "ISO date YYYY-MM-DD"
    },
    "assigned_to": {
      "type": "string"
    },
    "task_type": {
      "type": "string"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "entity_type",
    "entity_id",
    "subject",
    "due_date",
    "idempotency_key"
  ],
  "additionalProperties": false
}
set_candidate_tagsSchreiben

Replace manual candidate tags. Conflict if updated_at does not match.

Does not change pipeline auto-tags.

OAuth-Scopes
candidates:write
Plyce-Rechte
tags.candidates.edit
{
  "type": "object",
  "properties": {
    "candidate_id": {
      "type": "string"
    },
    "tags": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "updated_at": {
      "type": "string",
      "description": "Expected candidates.updated_at for conflict check"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "candidate_id",
    "tags",
    "idempotency_key"
  ],
  "additionalProperties": false
}
update_candidate_availabilitySchreiben

Set candidates.availability text. Conflict on updated_at.

Does not run status automation beyond normal DB triggers.

OAuth-Scopes
candidates:write
Plyce-Rechte
candidates.update
{
  "type": "object",
  "properties": {
    "candidate_id": {
      "type": "string"
    },
    "availability": {
      "type": "string"
    },
    "updated_at": {
      "type": "string"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "candidate_id",
    "availability",
    "idempotency_key"
  ],
  "additionalProperties": false
}
prepare_emailEntwurf

Store a client-written subject/body as a draft (MAIL-01: draft). Send only after a human confirms in plyce via email-send.

Does not call plyce-AI to write the mail. Does not send. Does not invoke email-send.

OAuth-Scopes
email:prepare
Plyce-Rechte
email.send
{
  "type": "object",
  "properties": {
    "to_entity_type": {
      "type": "string",
      "enum": [
        "candidate",
        "contact"
      ]
    },
    "to_entity_id": {
      "type": "string"
    },
    "to_email": {
      "type": "string"
    },
    "subject": {
      "type": "string",
      "description": "Written by the user model, not plyce AI"
    },
    "body_text": {
      "type": "string"
    },
    "body_html": {
      "type": "string"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "to_entity_type",
    "to_entity_id",
    "to_email",
    "subject",
    "idempotency_key"
  ],
  "additionalProperties": false
}
create_candidateSchreiben

Create a candidate (minimal fields). Returns id and plyce URL.

Does not upload CVs or run matching. For CV-based create use create_candidate_from_cv.

OAuth-Scopes
candidates:write
Plyce-Rechte
candidates.create
{
  "type": "object",
  "properties": {
    "first_name": {
      "type": "string"
    },
    "last_name": {
      "type": "string"
    },
    "email": {
      "type": "string",
      "description": "Required; unique per tenant (DB NOT NULL + unique)."
    },
    "phone_primary": {
      "type": "string"
    },
    "origin": {
      "type": "string"
    },
    "tags": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "first_name",
    "last_name",
    "email",
    "idempotency_key"
  ],
  "additionalProperties": false
}
create_companySchreiben

Create a company. Returns id and plyce URL.

Does not enrich from the web or invent website data.

OAuth-Scopes
companies:write
Plyce-Rechte
companies.create
{
  "type": "object",
  "properties": {
    "name": {
      "type": "string"
    },
    "website": {
      "type": "string"
    },
    "industry": {
      "type": "string"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "name",
    "idempotency_key"
  ],
  "additionalProperties": false
}
create_contactSchreiben

Create a contact linked to a company. Pass company_id (prefer search first) or rely on a unique email-domain match.

Does not create a company automatically. Ambiguous domain matches fail.

OAuth-Scopes
contacts:write
Plyce-Rechte
contacts.create
{
  "type": "object",
  "properties": {
    "first_name": {
      "type": "string"
    },
    "last_name": {
      "type": "string"
    },
    "email": {
      "type": "string"
    },
    "phone_primary": {
      "type": "string"
    },
    "company_id": {
      "type": "string",
      "description": "Required unless email domain uniquely matches one company (non-public domains). Prefer search_entities first."
    },
    "position": {
      "type": "string"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "first_name",
    "last_name",
    "idempotency_key"
  ],
  "additionalProperties": false
}
create_projectSchreiben

Create a project for an existing company (company_id required). Optional start_date, work_mode, contract type, tags, languages, notes and job text fields.

Does not invent job ads or publish to boards.

OAuth-Scopes
projects:write
Plyce-Rechte
projects.create
{
  "type": "object",
  "properties": {
    "title": {
      "type": "string"
    },
    "company_id": {
      "type": "string",
      "description": "Required company UUID"
    },
    "status": {
      "type": "string",
      "description": "Default Aktiv"
    },
    "start_date": {
      "type": "string",
      "description": "Projektstart (ISO date YYYY-MM-DD)"
    },
    "work_mode": {
      "type": "string",
      "enum": [
        "Vor Ort",
        "Hybrid",
        "Remote"
      ]
    },
    "anstellungsart": {
      "type": "string",
      "description": "Vertragstyp (tenant-specific label; trimmed, max ~100 chars)"
    },
    "tags": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "languages": {
      "type": "array",
      "items": {
        "type": "string"
      }
    },
    "description_rich": {
      "type": "string",
      "description": "Notes (UI Notes field)"
    },
    "requirements_rich": {
      "type": "string",
      "description": "Stellenbeschreibung"
    },
    "anschreiben": {
      "type": "string",
      "description": "Anschreiben / cover letter text"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "title",
    "company_id",
    "idempotency_key"
  ],
  "additionalProperties": false
}
add_candidate_to_projectSchreiben

Create an application (default status Long List). Fails on duplicates.

Does not send introduction emails.

OAuth-Scopes
candidates:write, projects:write
Plyce-Rechte
applications.create
{
  "type": "object",
  "properties": {
    "candidate_id": {
      "type": "string"
    },
    "project_id": {
      "type": "string"
    },
    "status": {
      "type": "string",
      "description": "Default Long List"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "candidate_id",
    "project_id",
    "idempotency_key"
  ],
  "additionalProperties": false
}
update_application_statusSchreiben

Move an application along allowed pipeline transitions (same rules as plyce UI).

Does not send Absage mail — use send_rejection for that.

OAuth-Scopes
candidates:write, projects:write
Plyce-Rechte
applications.update
{
  "type": "object",
  "properties": {
    "application_id": {
      "type": "string"
    },
    "status": {
      "type": "string",
      "description": "Target pipeline status (allowed transitions only)"
    },
    "updated_at": {
      "type": "string",
      "description": "Optional conflict check"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "application_id",
    "status",
    "idempotency_key"
  ],
  "additionalProperties": false
}
send_rejectionSchreiben

Set Absage status and optionally send the rejection email immediately (mailbox or platform_resend). Action-scoped send under mcp_writes_enabled — no separate email flag.

Not free-form mail; not delayed Absage (use plyce UI). Requires send_email true to send.

OAuth-Scopes
candidates:write, email:prepare
Plyce-Rechte
applications.update, email.send
{
  "type": "object",
  "properties": {
    "application_id": {
      "type": "string"
    },
    "absagegrund": {
      "type": "string"
    },
    "variant": {
      "type": "string",
      "enum": [
        "sie",
        "du",
        "en"
      ]
    },
    "send_email": {
      "type": "boolean",
      "description": "Must be true to send mail immediately; false = status only"
    },
    "new_status": {
      "type": "string",
      "description": "Absage von: Kunde|Agency|Kandidat (default Kunde)"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "application_id",
    "absagegrund",
    "variant",
    "send_email",
    "idempotency_key"
  ],
  "additionalProperties": false
}
prepare_document_uploadSchreiben

Create a short-lived signed upload URL for a candidate, company or project document. PUT the file, then call confirm_document_upload.

Does not accept file bytes in the tool call. Does not insert a CRM document row until confirm.

OAuth-Scopes
mcp
Plyce-Rechte
documents.upload
{
  "type": "object",
  "properties": {
    "type": {
      "type": "string",
      "enum": [
        "candidate",
        "company",
        "project"
      ]
    },
    "id": {
      "type": "string",
      "description": "Parent entity UUID. Omit with kind=cv to stage a CV before create_candidate_from_cv."
    },
    "file_name": {
      "type": "string"
    },
    "mime_type": {
      "type": "string"
    },
    "file_size_bytes": {
      "type": "integer",
      "description": "Declared size; must match uploaded bytes"
    },
    "document_type": {
      "type": "string",
      "description": "Candidate document type (default resume) or company/project document_type"
    },
    "kind": {
      "type": "string",
      "enum": [
        "document",
        "cv"
      ],
      "description": "cv allows image OCR types for candidates only; default document"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "type",
    "file_name",
    "mime_type",
    "file_size_bytes",
    "idempotency_key"
  ],
  "additionalProperties": false
}
confirm_document_uploadSchreiben

Validate the uploaded object (size + magic bytes) and insert the CRM document row. On failure the storage object is deleted.

Does not re-upload bytes; call prepare_document_upload first. Staging CVs use create_candidate_from_cv instead.

OAuth-Scopes
mcp
Plyce-Rechte
documents.upload
{
  "type": "object",
  "properties": {
    "upload_id": {
      "type": "string",
      "description": "From prepare_document_upload"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "upload_id",
    "idempotency_key"
  ],
  "additionalProperties": false
}
create_candidate_from_cvSchreiben

Finalize a staging CV upload: parse → analyze → create or update candidate by email + attach resume. Uses the AI CV analyzer with the caller JWT.

Not the minimal create_candidate field form. Does not accept raw file bytes (prepare + PUT first).

OAuth-Scopes
candidates:write
Plyce-Rechte
candidates.create, documents.upload
{
  "type": "object",
  "properties": {
    "upload_id": {
      "type": "string",
      "description": "Staging upload_id from prepare_document_upload(kind=cv, no id)"
    },
    "first_name": {
      "type": "string",
      "description": "Override when analyzer returns empty"
    },
    "last_name": {
      "type": "string",
      "description": "Override when analyzer returns empty"
    },
    "email": {
      "type": "string",
      "description": "Override when analyzer returns empty"
    },
    "phone_primary": {
      "type": "string"
    },
    "idempotency_key": {
      "type": "string"
    }
  },
  "required": [
    "upload_id",
    "idempotency_key"
  ],
  "additionalProperties": false
}
Agent-ready documentation

Für Menschen gestaltet.
Für KI lesbar.

Der Einstieg für KI-Clients: Authentifizierung, Tool-Schemas, Beziehungen und Regeln zum vollständigen Nachladen. Maschinenlesbare Dokumente werden mit dem Tool-Katalog erzeugt.

llms.txt ↗Agent guide ↗Tool schemas ↗Capabilities ↗
// Session
get_current_user_context({})

// Recruiter lists (one call)
list_projects({ "status": "Aktiv", "owner_id": "me" })
list_applications({
  "status": "Interview",
  "include": ["candidate", "project", "company"]
})

// Follow next_cursor. Prefer overviews over relation walks.
Gut zu wissen

Fragen zum Verbinden.

Warum sehe ich plyce in diesem Chat noch nicht?

Die Anmeldung gilt für den Client, in dem du sie eingerichtet hast. Prüfe dort, ob der Server aktiviert und OAuth abgeschlossen ist. Einige Clients laden die Werkzeuge erst in einer neuen Sitzung. Eine vorhandene CLI-Verbindung aktiviert keine separate ChatGPT-App-Verbindung.

Neue Tools fehlen trotz erfolgreicher Anmeldung?

ChatGPT kann einen gespeicherten Stand der App-Werkzeuge und Eingabefelder verwenden. Ein Deployment oder eine erneute OAuth-Anmeldung aktualisiert diesen Stand nicht automatisch. Öffne ChatGPT im Browser und prüfe als Admin die Workspace-Einstellungen → Apps. Bei Enterprise/Edu: Aktionen aktualisieren (Refresh), neue Aktionen prüfen und freigeben. Für bereits veröffentlichte Business-Apps ist laut OpenAI derzeit erneutes Erstellen und Veröffentlichen nötig. Danach einen neuen Chat öffnen und die aktualisierte App auswählen. Aktuelle OpenAI-Anleitung.

Zur Diagnose liefert get_current_user_context die Katalogrevision und die serverseitig erlaubten Tools (allowed_tools / discovery.tool_count). Fehlt ein dort genanntes Tool im Chat — typisch list_entity_documents und read_document bei 36 serverseitig vs. 34 im Host — ist der Host-Katalog veraltet und das Tool noch nicht getestet. Dokument-Tools brauchen zusätzlich mcp.documents.read (und mcp.content.read für read_document). Die alte search-Suche ersetzt keine strukturierte Kandidatensuche. Eine öffentliche Tool-Liste bestätigt weder deine Berechtigungen noch erfolgreiche Datenabfragen.

Was bedeutet eine 401-Antwort?

Ein erster nicht authentifizierter Aufruf kann mit 401 die OAuth-Anmeldung auslösen. Wiederholte 401 nach der Anmeldung sind ein Fehler: Anmeldung erneuern und die Servermeldung prüfen lassen. Eine erfolgreiche OAuth-Anmeldung allein bestätigt noch keinen erfolgreichen CRM-Abruf.

Liest meine KI automatisch die gesamte Datenbank?

Die KI ruft gezielt verfügbare Werkzeuge auf. Listen werden seitenweise ausgegeben; zusätzliche Beziehungen und Dokumentabschnitte werden gesondert nachgeladen. Sichtbar sind nur die für deinen Benutzer freigegebenen Inhalte.

Wie vollständig ist die Radius-Suche?

Die Suche verwendet die Luftlinie vom angegebenen Suchzentrum. Profile ohne Koordinaten werden nicht als Radius-Treffer ausgegeben; ihre Anzahl wird separat gemeldet. Bei mehrdeutigen Orten sind präzisere Ortsangaben oder Koordinaten nötig.

Welche Grenzen gibt es bei Dokumenten?

Die Textextraktion unterstützt PDF, DOCX und UTF-8-Text bis 10 MiB; PDFs bis 500 Seiten. Gescannte Seiten ohne Textebene benötigen OCR und werden entsprechend gekennzeichnet. Der Zugriff auf ältere Dateipfade kann eine Bereinigung der Storage-Zuordnung erfordern.

Welche Daten erhält mein KI-Anbieter?

Die Ergebnisse der aufgerufenen Werkzeuge werden an deinen verbundenen KI-Client übermittelt. Dessen Datenverarbeitung und Aufbewahrung richten sich nach deinem Anbieter und Vertrag. Gib deshalb nur die benötigten Inhalte und Berechtigungen frei.

Suche & Reporting

Die richtige Person.
Der passende Zeitraum.

Menschen über Namen finden, Zusammenhänge nachvollziehen und Zahlen mit klarer Bedeutung abrufen.

01 / Menschen

Namen statt UUID-Rätsel.

„Was hat Malin im August gemacht?“ Der MCP löst Namen und Benutzer-IDs innerhalb deines Tenants auf. Mehrere Treffer? Deine KI fragt nach. Auch inaktive Mitarbeiter bleiben für historische Berichte auflösbar.

resolve_user({ "name": "Malin" })
02 / Zeiträume

Vom Kalendertag bis zum Quartal.

Eigene Von-/Bis-Daten, letzte Kalenderwoche, letzter Monat, Quartale und laufendes Jahr – einheitlich in Europe/Berlin, inklusive Sommerzeit. Tagesangaben schließen den Bis-Tag ein.

from: "2026-08-01", to: "2026-08-31"
03 / Ähnlichkeit

SPS findet auch Simatic S7.

Offengelegte verwandte Skill-Begriffe erweitern die Suche. Ähnliche Kandidaten erhalten einen regelbasierten Vergleich: Skills 60 %, Jobtitel 25 %, Seniorität 15 %. Fehlende Informationen und die Größe des geprüften Kandidatenpools bleiben sichtbar.

Keine Einstellungswahrscheinlichkeit und keine KI-Prognose. Standort und Status filtern den Suchraum.

Historie mit Belegen.

Sent Outs, Interviews, Offers und Placements aus protokollierten Statuswechseln. Historische Ereignisse und aktuelle Pipeline-Bestände sind ausdrücklich getrennt. Nicht auflösbare Beziehungen werden ausgewiesen.

Dealwerte mit klarer Zuordnung.

Placement-Dealwerte nach Recruiter-Split, Kunde, Projekt oder Placement – nur mit Finanzberechtigung. Fehlende Werte und Währungen bleiben erkennbar. Das ist kein fakturierter Umsatz und keine realisierte Marge.

Aktivitäten richtig einordnen.

Manuell, automatisiert, System, Bot, Bulk oder unbekannt. Nachweisbare Automationen lassen sich ausschließen. Unbekannte Herkunft bleibt standardmäßig enthalten und wird im Bericht gezählt.

Bei Listen jede Folgeseite nachladen. Ein Benutzerbericht kann KPI-Zahlen und einen zusätzlich berechtigungsgeprüften Aktivitäten-Drilldown liefern. Zeitreihen zählen Aktivitäten, keine Arbeitsstunden.