Servidor MCP de viAct

• Endereçohttps://mcp.viact.ai/mcpfora do registro oficial: respondeu quando perguntado

Nenhuma entrada do registro oficial de MCP nomeia este servidor. Ele respondeu em viact.ai quando o Atlas perguntou no endereço onde servidores costumam ficar, então é do próprio fornecedor. Clientes e catálogos que listam servidores a partir do registro não o mostram.

• Estadorespondeverificado em 2026-10-27
• Ferramentas87
• O servidor se chamadot-ai-backend
• Versão do servidor2.14.4
• Versão do protocolo2025-06-18

Em 2026-10-27 o servidor respondeu a initialize e tools/list sem login. Primeira resposta em 298 ms.

• Ainda não avaliamos se as ferramentas fazem o trabalho do produto.
• Verificado em 1 dos 7 dias necessários para julgar a confiabilidade.

Dito por regras a partir das medições acima. A primeira chamada provável é a escolha de um modelo na lista de ferramentas; leia a descrição antes de confiar nela.

• Nota INITE MCPdepois de uma semanaa nota precisa de 7 verificações diárias: 1 até agora
• Respondeumedindo1 das 7 verificações diárias necessárias para mostrar a porcentagem
• Mediana da primeira resposta684 ms
• Espaço no contexto do agente≈ 13.016 tokensas definições de todas as ferramentas (87), que o agente carrega a cada turno
• Definições das ferramentasC · 69 de 100
• Resultados tipados4 de 87 ferramentasdeclaram outputSchema: o resultado chega tipado e é lido de uma vez

• Parâmetros sem tipo: em 7 de 87
• Parâmetros sem descrição: em 69 de 87
• Sem dizer se só lê ou se pode destruir: 87 de 87

Um programa verifica o servidor uma vez por dia: initialize e tools/list, sem modelo. Os tokens são o tamanho das definições, a cerca de quatro caracteres por token. A nota é a média de seis verificações do nome, da descrição, do esquema e das anotações de cada ferramenta; diz se as ferramentas estão bem descritas, não se funcionam. A nota pesa só o que foi medido: se responde (25%), se faz o trabalho do produto (25%), definições (20%), contexto (10%), uso (10%) e a facilidade de conectar (10%). Uma parte não medida fica de fora, e uma instrução oculta limita a nota a 40. O estado dos servidores MCP dos fornecedores: método e números →

• company-getCompany

  [Company] Executes GET /api/company/{id}

• groupingUnit-createGroupingUnit

  [Grouping Unit] Creates a new grouping unit in the hierarchy

• groupingUnit-getGroupingUnitTree

  [Grouping Unit] Returns the hierarchical structure of grouping units starting from the specified unit **Path Parameters:** - **unitId**: Grouping Unit ID

• groupingUnit-getRootUnits

  [Grouping Unit] Returns all active root-level grouping units

• groupingUnit-getGroupingLevels

  [Grouping Unit] Returns the available grouping levels ordered by hierarchy

• groupingUnit-addCompaniesToUnit

  [Grouping Unit] Adds one or more companies to the specified grouping unit and propagates the assignment up the hierarchy **Path Parameters:** - **id**: Grouping Unit ID

• monitor-getLists

  [Monitor] Executes GET /api/monitor

• monitor-getMonitor

  [Monitor] Executes GET /api/monitor/{id}

• stats-getObjectCountingStatistic

  [Stats] Get Object counting statistic

• stats-getSmartWatchStats

  [Stats] Get Critical incident statistic

• stats-getSWActivityStats

  [Stats] Get Smart watch activity statistic

• stats-getGasSensorStats

  [Stats] Get gas sensor statistic

• stats-getAlertsTrendSummary

  [Stats] Get trend change alerts

• stats-getAllAlertsTrendSummary

  [Stats] Get trend change alerts

• stats-getAlertTrendByRange

  [Stats] Get alert trends summary by range and duration

• stats-getAlertSummaryByLatLng

  [Stats] Get alerts summary by lat and long

• stats-getAllAlertTrends

  [Stats] Get trend change alerts

• detection-getAllDetections

  [Detection] Executes GET /api/detection

• engines-searchEngines

  [Engines] Executes GET /api/engines/search

• inventory-getAllInventory

  [Inventory] Executes GET /api/inventory

• inventory-getInventoryStatsValidity

  [Inventory] Executes GET /api/inventory/stats-validity

• inventory-getInventory

  [Inventory] Executes GET /api/inventory/{id}

• companyMonitors-createNewCME

  [Company Monitors] Executes POST /api/companies-monitors-engines

• smartLock-getSmartLockEventLogs

  [Smart Lock] Executes GET /api/smart-lock/events

• smartLock-getAllSmartLockDevices

  [Smart Lock] Executes GET /api/smart-lock/devices

• smartLock-getSmartLockDevice

  [Smart Lock] Executes GET /api/smart-lock/devices/{id}

• permit-getPermitStatsStatus

  [Permit] Executes GET /api/permit/stats/status

• permit-listPermits

  [Permit] Executes GET /api/permit

• permit-createPermitDraft

  [Permit] Executes POST /api/permit/draft

• permit-duplicatePermit

  [Permit] Executes POST /api/permit/{id}/duplicate

• permit-getPermitById

  [Permit] Executes GET /api/permit/{id}

• permit-updatePermit

  [Permit] Executes PUT /api/permit/{id}

• permit-getPermitLogs

  [Permit] Executes GET /api/permit/{id}/logs

• permit-publishPermit

  [Permit] Executes PUT /api/permit/{id}/publish

• permit-listPermitTemplates

  [Permit] Executes GET /api/permit-template

• permit-listPermitCategories

  [Permit] Executes GET /api/permit-template/category

• permit-getPermitTemplateById

  [Permit] Executes GET /api/permit-template/{id}

• gasMeter-getGasDeviceMonitoring

  [Gas Meter] Executes GET /api/gas-meter/{deviceId}/device-monitoring

• gasDevice-getGasDevices

  [Gas Device] Executes GET /api/gas-devices

• gasDevice-getGasDeviceAlerts

  [Gas Device] Executes GET /api/gas-devices/alerts

• gasDevice-getGasDevice

  [Gas Device] Executes GET /api/gas-devices/{id}

• droneDetection-getAllMonitorVideos

  [Drone Detection] Retrieve list of monitor videos with filters for text search, pagination, and engines

• droneDetection-getDetectionsByDrone

  [Drone Detection] Retrieves drone detections filtered by company_code and/or monitor_id (drone). company_code defaults to the authenticated user's company. Supports optional period and status filters. Executes as a single query via…

• droneDetection-getMonitorVideo

  [Drone Detection] Retrieve a specific monitor video by its monitor ID including monitor details and detection count

• cpecs-listPaymentPackages

  [CPECS] Triage view of payment-application packages for the active project (resolved from request context). Supports status / search / pagination. Read-only.

• cpecs-createPaymentPackage

  [CPECS] Create a package end-to-end in the active project (from request context): create → save cost-component groups → save line items → optionally publish. Confirm-gated in the agent runtime; the agent shows the draft for approval first.

• cpecs-getPackageOverview

  [CPECS] Single rich read for a package: KPIs, grouped totals, duplicate flags and status. Replaces the multi-call fan-out the UI does.

• cpecs-getPackageCostStructure

  [CPECS] Cost-component groups joined with their line items — "what is this package made of?". Grounds mapping & over-claim checks.

• cpecs-listPackagePaymentApps

  [CPECS] All payment applications (bids) submitted under a package and their workflow state — feeds vetting selection.

• cpecs-getPaymentAppReviewBundle

  [CPECS] Core vetting read. One call returns the full payment app (header + items by bucket + totals), prior-cert cross-check deltas, and the audit trail.

• cpecs-getItemEvidence

  [CPECS] Prior-period diff, change history and the note thread for one item — used to justify a reject. Lazy-loaded so the review bundle stays lean.

• cpecs-benchmarkLineItems

  [CPECS] The agent's pricing evidence. Batch-compares the supplied line items (pass them from get_payment_app_review_bundle) against the Rate DB and returns, per item, the matched band (p25/median/p75), Δ vs median %, escalation flag and a…

• cpecs-searchRateDb

  [CPECS] Free-text / filtered similarity search of the Rate Database — ad-hoc "find me rates for X" and grounding for the wizard.

• cpecs-getRateBenchmark

  [CPECS] Single-item benchmark with optional time-escalation — "is this unit rate fair vs market?" for one description.

• cpecs-listSoccCatalogue

  [CPECS] Valid SoCC (Standard of Cost Components) codes for the active project (from request context) — grounding so SoCC mappings use real codes, not hallucinated ones. Read-only.

• cpecs-classifyItem

  [CPECS] All reversible edits to one invoice line, behind an action discriminator: "assess" (accept), "map_socc" (set SoCC code), "quick_add" (create + map a cost-component line), or "note" (append a note). Returns the refreshed payment app.

• cpecs-rejectItem

  [CPECS] Reject a line with a required reason (lands in the note thread + audit). mode="disallow" strikes it; mode="return" sends it back for backup. Confirm-gated in the agent runtime.

• cpecs-setReviewState

  [CPECS] Toggle the application review state. state="reviewed" promotes it (precondition: no items left unmapped/unassessed); state="in_review" reopens it. Returns the refreshed payment app.

• cpecs-certifyPaymentApp

  [CPECS] Certify the payment application — the money-moving final step. Irreversible. The agent must present the certify total + disallowed summary and obtain explicit human approval before calling this.

• cpecs-startSpecExtraction

  [CPECS] Enqueue an extraction job for a spec PDF already uploaded to S3 (presigned). Returns a job view; poll get_spec_extraction_status until done. Produces a DRAFT only — nothing is persisted to a package.

As primeiras 60 de 87, na ordem do servidor.