How much does LuMay cost?

Per agent SaaS starting prices: Voice Agents $249 a month, Translation $499, Insight, Forecasting and Anomaly Detection $999 each, Workflow Automation $1,499, Lexintis legal agents $7,500, and OCG and QMS Compliance $10,000 each, with the regulated agents enterprise led. Customer cloud adds about 25% and on premises about 40% plus a one time setup, both as annual subscriptions. Model usage is included within plan volumes, and the exact monthly price is confirmed on a call. Voxentis voice plans: Basic $249 with 500 minutes, Starter $499 with 1,000, Growth $999 with 2,000, overage $0.21 to $0.23 a minute, connectors $0.10 per active minute, Enterprise custom, 14 day free trial. Custom solutions start at a $25,000 pilot, and managed optimization runs $5,000 to $25,000 a month.

Does LuMay support MCP and tool calling?

Full under the Agentic Index verification standard. LuMay publishes public evidence of Model Context Protocol (MCP) support, tool calling, and API or SDK extensibility for connecting agents to external tools. Evidence basis: A REST API covers agents (create, update, clone, deploy), calls (status, transfer, end, transcript, recording), speech to text and text to speech, tools, call logs with export, metrics and pipeline configuration, and can generate a conversa. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does LuMay have SOC 2, SSO and role based access control?

Full under the Agentic Index verification standard. LuMay publishes public evidence of enterprise security and identity controls, covering a compliance posture such as SOC 2 or ISO 27001, single sign on (SSO or SAML), and role based access control (RBAC). Evidence basis: Sign on runs through Keycloak or the customer's own identity provider over SAML or OIDC, with JWT bearer tokens checked on every API call and role checks on every route. Tenant isolation is enforced by PostgreSQL row level security in the d. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Can LuMay be self hosted or run in a private VPC?

Full under the Agentic Index verification standard. LuMay publishes public evidence of self hosted or private VPC deployment, along with data residency options for controlling where agent data is processed and stored. Evidence basis: The same Docker images run in five models: LuMay managed SaaS in its Azure tenancy, customer cloud inside the customer's Azure or AWS subscription, a private VPC with network isolation, on premises Kubernetes with full air gap capability, a. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Back to vendors

LuMay

Also known as: Lumay, LuMay AI, lumay.ai, LuMay Inc, Voxentis, Lexintis

Visit site

Agent builderindependentVerified 2026-10-18

Entry priceVoice Agents from $249 a month on SaaS; other agents from $499 to $10,000 a monthFull pricing detail

LuMay is an enterprise agent platform whose single runtime, orchestration engine, connector layer and governance layer sit under packaged agents for phone calls (Voxentis), legal work (Lexintis), knowledge search, workflow automation, forecasting, anomaly detection and translation. Flows are built visually or from a plain English description, with approval gates, versioning and delegation between agents, and the same containers run as SaaS, in the customer's Azure or AWS account, on premises or air gapped.

LuMay, Inc. sells what it calls an AI agent factory for regulated enterprises in legal, healthcare, financial services, technology and manufacturing: one platform for building and governing agents, plus packaged agents priced one by one. AI development, integration and training services run as a separate line.

Every agent runs on one Core Engine that retrieves context from a knowledge base, reasons with OpenAI or Anthropic models, and routes tool calls through a registry to connectors for HubSpot, Salesforce, Freshdesk, Zendesk, ServiceNow, SharePoint, email, SMS, SQL databases and any MCP server. The Orchestration Engine holds the business logic: flows authored in a visual editor or generated by a plain English builder, with branching, loops, tool call nodes, approval gates per node, retries, delegation between agents, and version history with diffs and one click rollback.

Voxentis voice agents run on a Pipecat pipeline with Deepgram transcription and ElevenLabs or Deepgram voices, answer and place calls over Twilio, look up records and create tickets mid call, and transfer to people with a summary. Insight Agents answer across SharePoint, Confluence and Jira with citations and per user permissions, and Lexintis bundles six legal agents for intake, citation checks, billing readiness, finance questions and outside counsel guideline compliance. Each exchange is traced with OpenTelemetry.

Published SaaS starting prices run from $249 a month for Voice Agents and $499 for Translation to $999 for Insight, Forecasting and Anomaly Detection, $1,499 for Workflow Automation, $7,500 for Lexintis and $10,000 for OCG and QMS Compliance. Customer cloud adds about 25% and on premises about 40% plus setup, model usage is included, and the exact price is confirmed on a call; Voxentis plans add included minutes and a 14 day free trial.

Security covers Keycloak single sign on over SAML or OIDC, role based access, tenant isolation in the database, PII masking, Azure Key Vault and audit trails. The site footer claims SOC 2 Type II and ISO 27001 certification, while the security page speaks of alignment with SOC 2, GDPR and HIPAA, so buyers should expect to request the reports.

LuMay fits regulated mid sized and large organizations that want governed agents for calls, legal billing, knowledge and workflows in their own cloud or data center, with implementation help. It fits less well for teams that want to try a builder on their own without a sales conversation, need browser or desktop automation, or want a built in evaluation suite.

Vendor details

Canonical URL

https://www.lumay.ai

Subcategory

Enterprise agent platform with a provider agnostic runtime, governed orchestration and packaged voice, legal, knowledge and workflow agents, deployable from SaaS to air gapped

Funding status

Privately held LuMay, Inc., based in Dover, Delaware; no outside funding disclosed.

Company status

independent

Use cases & customers

Primary use cases

inbound reception and outbound calling with record lookups and bookings during the callcited answers over SharePoint, Confluence and Jira for employeesmultistep approval workflows across ServiceNow, SAP and ERP systemslegal invoice readiness, OCG compliance and contract reviewdocument and website translationforecasting and anomaly detection over operational data

Target customers

regulated enterprises in legal, healthcare, financial services and manufacturinglaw firms and corporate legal departments managing billing and outside counsel guidelinescontact centers and service businesses automating inbound and outbound callsIT and operations teams connecting ServiceNow, SAP and ERP workflowsorganizations that must run AI in their own cloud, on premises or air gapped

Deployment options

LuMay managed SaaS on AzureCustomer cloud in Azure or AWSPrivate cloud or VPCOn premises Kubernetes, including air gappedHybrid with an on premises data plane

Integrations

Connectors for HubSpot, Salesforce, Zoho, Freshdesk, Zendesk, Zoho Desk, ServiceNow, SharePoint, OneDrive, Google Drive, Google Sheets, Google Calendar, Square Appointments, Shopify, Stripe, SendGrid, Mailgun, Resend, Twilio voice and SMS, and SQL databases, plus SAP, Microsoft 365 and data warehouses through workflow adapters, and Aderant, Elite 3E and ebilling platforms for legal agents. Any MCP server, REST API functions and webhooks extend the set. Models come from OpenAI, Anthropic and Google, and speech from Deepgram, ElevenLabs, Azure, Google and PlayHT.

In practice

A multi location clinic points its after hours Twilio line at a Voxentis voice agent that recognizes returning patients, checks availability in Google Calendar, books the appointment during the call and texts a confirmation, transferring urgent callers to staff with a summary of the conversation.

A law firm's billing team runs BillingGuard against Elite 3E before invoices go out and sees which ones fail client guidelines, how much revenue is exposed and who needs to act, with human review before anything reaches the client.

A bank's operations group deploys Workflow Automation Agents inside its own Azure subscription to move loan applications through document checks, ServiceNow tasks and SAP updates, with approvals routed by value threshold and every step traced.

Sources & related URLs

Related / legacy domains

https://www.lumay.ai/ai-platform https://www.lumay.ai/pricing https://www.lumay.ai/ai-products/voice-agent https://www.lumay.ai/docs/docs-api-overview

Research sources

Agentic Index coverage score

12.0 / 14 capabilities · 86%

Integrations & Tool Calling	Full
When a model emits a tool call, a registry routes it by tool name and provider to a connector instance that authenticates with an OAuth token or a vaulted API key, calls the outside system and returns a normalized result. Connector categories share tool names across vendors: HubSpot and Salesforce for sales records, Freshdesk, Zendesk, Zoho Desk and ServiceNow for tickets, Resend, SendGrid and Mailgun for email, SharePoint and OneDrive for documents, Google Sheets and SQL databases for identity checks, and Twilio for SMS. Teams can also register API functions and any MCP server, and swapping Freshdesk for Zendesk leaves the agent's flow untouched. Sourcelumay.ai/ai-platform/connectorsread 2026-10-18
Workflow Orchestration	Full
Flows are authored in a visual editor with message, input, condition, tool call and transfer nodes, branching, loops and variables, or generated by a ten step plain English builder that wires up the connectors and can be edited afterward. At run time the flow engine evaluates each node, dispatches tool calls, enforces approval gates and handles retries and compensation when a step fails. One agent can delegate subtasks to another, and Workflow Automation Agents chain a dozen or more agent and system steps across ServiceNow, SAP and ERP systems, built on LangGraph, AutoGen and CrewAI. Sourcelumay.ai/ai-platform/orchestration-engineread 2026-10-18
Knowledge Grounding & RAG	Full
Documents uploaded to a knowledge base vault as PDF or DOCX are indexed for vector search, and relevant passages are retrieved before every model call. Insight Agents read SharePoint, Confluence, Jira and document management systems in place, keep each user's permissions, answer with inline citations naming the source document and its version, and take synonyms, glossaries and document priorities so internal jargon resolves. Voice plans include knowledge base storage from 1 MB on Starter to 1 GB on Enterprise, with more at $20 per 200 MB a month, and retrieval carries no per query charge. Sourcelumay.ai/ai-products/insight-agentread 2026-10-18
Human Oversight & Guardrails	Full
Approval gates are set per flow node and enforced in the orchestration layer rather than in application code, so an action above a configured risk threshold waits for human review. Workflow Automation Agents route approvals by policy, using value thresholds, risk scores or jurisdiction flags, and people approve only the steps flagged as needing judgment, each with full context. Voice agents escalate on rules or rising frustration and transfer warm, cold or blind to a person or another agent with a live summary, and Lexintis legal agents keep human review at every step before output reaches a client. Sourcelumay.ai/ai-platform/orchestration-engineread 2026-10-18
Security, Identity & Governance	Full
Sign on runs through Keycloak or the customer's own identity provider over SAML or OIDC, with JWT bearer tokens checked on every API call and role checks on every route. Tenant isolation is enforced by PostgreSQL row level security in the database itself, PII is detected and masked before it reaches the model, and secrets sit in Azure Key Vault with workload identity. The Voxentis API adds viewer, operator, builder, admin and owner roles, groups scoped to named agents, permission scoped API keys and an audit log of logins, key use and changes. On certification LuMay's own pages disagree: the site footer says SOC 2 Type II and ISO 27001 certified, while the security page describes SOC 2, GDPR and HIPAA as alignment and links no report. HIPAA work runs in private cloud or on premises under a BAA. Sourcelumay.ai/ai-platform/securityread 2026-10-18
Observability & Auditability	Full
Every request, database query, queue message and connector call emits an OpenTelemetry span tied to a correlation ID, and a collector ships traces to Grafana Tempo, metrics to Prometheus and logs to Loki in every deployment model. Call logs hold an event timeline for each call, from answer and transcription through tool invocations and speech to hang up, along with the transcript, recording, summary, intent, topics and sentiment, and can be filtered on more than 20 fields or exported as CSV, JSON or Excel. Dashboards track containment, escalation, connector success, sentiment, call volume and cost per outcome. Sourcelumay.ai/ai-platform/analyticsread 2026-10-18
Memory & State Persistence	Partial
Conversation history persists across turns within a session, flows collect caller answers into named variables, and longer context sits in a call session store. Voice agents recognize returning callers by phone number or a verification step and pull their history and preferences before the conversation starts, and Insight Agents handle follow up questions within a session. Beyond that, memory means the records and transcripts written to connected systems, since there is no store of learned facts that an agent updates and that persists across separate conversations under its own controls. Sourcelumay.ai/ai-platform/core-engineread 2026-10-18
Deployment & Data Residency	Full
The same Docker images run in five models: LuMay managed SaaS in its Azure tenancy, customer cloud inside the customer's Azure or AWS subscription, a private VPC with network isolation, on premises Kubernetes with full air gap capability, and hybrid, with the management plane in the cloud and the data plane on premises. Governance, observability and the agent runtime stay identical across them, Azure Container Apps serve SaaS and customer cloud, and any Kubernetes distribution runs the rest. Air gapped installs get a custom architecture and an offline update process, and customer cloud and on premises run as annual subscriptions. Sourcelumay.ai/ai-platform/deploymentread 2026-10-18
Prebuilt Agents / Templates / Packs	Full
Packaged agents cover phone calls (Voxentis), translation, knowledge search (Insight), forecasting, anomaly detection, workflow automation, and regulated work in legal, OCG and QMS compliance. Lexintis bundles six legal agents: Intake Flow for personal injury intake, Virtual Paralegal for citation checks, BillingGuard for invoice readiness, BillingOps and Metrics Insight for questions over 3E and finance data, and OCG Compliance for outside counsel guidelines. Voice agents start from conversation flow templates managed through the API, and industry editions target healthcare, financial services and manufacturing. Sourcelumay.ai/pricingread 2026-10-18
Triggers & Channel Coverage	Full
Agents are reached through voice, chat, Microsoft Teams, a web portal and a REST API, and Insight Agents also answer in Slack. Voxentis voice agents answer inbound calls on Twilio numbers and place outbound calls from a scheduler, including batch campaigns for reminders, collections and lead qualification, with SMS or WhatsApp messages sent during or after a call. Calls and workflows can be started from system events, webhooks, workflow tools and API requests, and Workflow Automation Agents add event driven and scheduled runs for approvals, document routing and data pipelines. Sourcelumay.ai/ai-platform/architectureread 2026-10-18
Model Flexibility & Routing	Full
The reasoning runtime is provider agnostic, so an agent can move between OpenAI and Anthropic models, or any compatible API, without changing its flow, prompts or connectors. Voice agent configuration takes OpenAI GPT models, Anthropic Claude Sonnet and Haiku, Google Gemini Pro or a custom fine tuned model, set per agent along with temperature and token limits, and speech runs on Deepgram, Google, Azure or Whisper for transcription and ElevenLabs, Azure, Google or PlayHT for voices. Token budgets and latency caps apply per agent, and Enterprise can deploy a private or fine tuned model inside the platform. Sourcelumay.ai/ai-platform/core-engineread 2026-10-18
APIs / SDKs / MCP Extensibility	Full
A REST API covers agents (create, update, clone, deploy), calls (status, transfer, end, transcript, recording), speech to text and text to speech, tools, call logs with export, metrics and pipeline configuration, and can generate a conversation flow from a plain English description. WebSockets stream live call audio and events, transcription and synthesis. MCP servers over stdio or HTTP register through the API with automatic tool discovery and are assigned to agents in their configuration, API tools wrap outside endpoints, and keys carry scoped permissions. Webhooks deliver call events where the plan includes them, with polling of call sessions as the fallback. Sourcelumay.ai/docs/docs-api-overviewread 2026-10-18
Testing, Debugging & Optimization	Partial
Each tool can be test executed through the API before an agent uses it, voices can be auditioned with sample text, and builders listen back to calls to tighten prompts and guardrails. Flow versions are tracked with a full diff between any two, the deployed version is locked, rollback takes one click, and agents support A/B testing across versions. Analytics add quality scores, sentiment and escalation rates to show where a flow is failing. There is no suite of simulated conversations or scored test sets to run before a version ships, and user acceptance testing comes as part of the implementation service. Sourcelumay.ai/ai-platform/orchestration-engineread 2026-10-18
Browser / Computer-use	Not documented
Agents act through connectors, API tools and MCP servers. No agent opens web pages, fills forms, clicks through an interface or reads a screen, and LuMay positions its platform against screen based RPA tools such as UiPath and Blue Prism. Work in systems without an API goes through custom connectors, scheduled file exchange or the integration services team. Voice agents send SMS or WhatsApp messages and transfer calls, and workflow agents move items through ServiceNow, SAP and ERP adapters, all through programmatic interfaces. Sourcelumay.ai/ai-platform/connectorsread 2026-10-18

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Pricing

Voice Agents from $249 a month on SaaS; other agents from $499 to $10,000 a month

Per agent monthly subscription with a published starting price, sized by each agent's volume driver (minutes, pages, data sources, workflows, invoices), with a deployment uplift for customer cloud or on premises

Trial available

Included quota

Voice Basic: 500 minutes; Starter: 1,000 minutes and 1 connector; Growth: 2,000 minutes, 3 connectors, 30 day analytics and MCP; Enterprise: 5,000+ minutes. Knowledge base storage 1 MB on Starter, 10 MB on Growth, 1 GB on Enterprise. Model usage included within published volumes.

What is public

Starting price for every packaged agent on SaaS, deployment uplifts, voice plan prices with included minutes, overage and connector rates, knowledge base storage limits, and services and training prices.

Billing mechanics

Per agent subscription, monthly or annual, with usage overage on voice; customer cloud and on premises run as annual subscriptions with support and upgrade governance.

Cost watchouts

Connector minutes on voice calls stack per connector, Twilio SMS carrier fees pass through, on premises adds a one time setup, and custom work starts at a $25,000 pilot.

Variable cost rationale

Starting prices are fixed per agent, but voice bills overage per minute and each live connector per active minute, deployment outside SaaS adds 25% to 40%, and the final price depends on volumes, integrations and security model confirmed on a call.

Additional watchouts

Voice connector charges and deployment uplifts can move the bill well above the starting price.

Overage / add-ons

Voice minutes over plan cost $0.23 on Basic, $0.22 on Starter and $0.21 on Growth; each connector bills $0.10 per active minute while engaged on a call; extra knowledge base storage $20 per 200 MB a month on Enterprise; Twilio SMS carrier fees pass through.

Sales call required

Mixed (some tiers require a call)

Free / trial

14 day free trial on voice plans, no credit card; proof of concept pilots for enterprise

Lowest paid plan

Voice Agents (Voxentis) Basic at $249 a month with 500 included minutes and $0.23 a minute overage

Commercial notes

Bundles of several agents go on one governed contract with volume pricing; regulated legal, OCG and QMS agents are enterprise led; AI development, managed optimization and Academy training are sold separately.

Key ambiguities

Every figure is a starting price confirmed on a call, the annual discount is two months free on one pricing page and 10% off voice base fees on another, and starting prices for agents other than voice do not show their included volumes.

Support SLA / resale

99.99% SLA with service credits on Voice Enterprise; dedicated CSM and white labeling available.

Missing data

Included volumes for non voice agents, enterprise voice pricing, and contract terms.

Agentic Index verified 2026-10-18

Official pricing page Docs

Data confidence: medium