# LuMay — Enterprise agentic AI platform with packaged voice, legal, compliance and workflow agents; SaaS, on-prem or air-gapped.

> Source: The Agents Index — https://theagentsindex.com/lumay (structured, researched, re-verified)
> Facts last verified: 2026-09-09

LuMay is an enterprise agentic AI platform whose packaged agents cover voice, legal, compliance, translation, forecasting, anomaly detection and workflow automation on one governed runtime. The same containers run as LuMay-hosted SaaS, inside your own cloud, on-premises or air-gapped.

| Fact | Value |
| --- | --- |
| Website | https://www.lumay.ai |
| Price | from $249/month billed monthly, $208/month billed annually |
| Free trial | Partly |

14-day free trial with no credit card on the Voxentis voice plans, giving full access to the chosen plan's features for the trial period. There is no standing $0 tier, and no trial is published for the other agents, which start at $499/mo.

## What it does

**Done-for-you service** · Yes — LuMay sells its own delivery people alongside the software: a discovery sprint, a fixed-scope production pilot and a dedicated monthly engineering pod, with priced ongoing managed optimisation.

> A dedicated team - architect, AI engineer, integration engineer, QA - for ongoing build, adoption, and optimization.
> — https://www.lumay.ai/pricing · Section 06 - Engineering & Services, 'AI Engineering Pod' card ($35K-$75K/month)

## How much it does unattended

**Runs autonomously** · Partly — The vendor positions its agents as deliberately semi-autonomous: they plan and act, but sensitive steps stop at a human approval gate rather than completing unattended.

> AI agents go further — interpreting a goal, planning the steps to complete it, retrieving information, taking approved actions, and checking the result.
> — https://www.lumay.ai/ai-agent-development-company · 'Beyond Chatbots' section, 'Goal Interpretation & Execution' paragraph; the comparison matrix below rates Autonomy as 'Controlled/semi-autonomous'

**Multi-agent** · Yes — Specialised agents run under a supervisor with delegation, shared state and conflict resolution, and the workflow product names LangGraph, Autogen and CrewAI as the orchestration underneath.

> For complex workflows, we design specialized agents (collection, analysis, validation) coordinated by a supervisor agent, with task delegation, shared state, conflict resolution, and human escalation
> — https://www.lumay.ai/ai-agent-development-company · 'Our AI Agent Development Services', item 05 'Multi-Agent Systems'

**Agent permissions** · Yes — Human-in-the-loop gates sit in the orchestration layer, so an agent's reach before it must ask is a configured policy rather than a fixed product behaviour.

> Human approvals for sensitive actions at orchestration layer.
> — https://www.lumay.ai/ai-platform/security · Security Controls table, 'HITL gates' row

## Which models it runs on

**Claude** · Yes — The runtime names Anthropic as a swappable provider alongside OpenAI; the vendor names the provider, not a specific Claude model version.

> Swap LLM providers (OpenAI ↔ Anthropic) without changing the agent, flow, or connectors.
> — https://www.lumay.ai/ai-platform/core-engine · Key Features table, 'Provider abstraction' row

**GPT** · Yes — OpenAI is one of the two named providers the reasoning loop calls, and the hosted default routes through Azure OpenAI inside Microsoft's enterprise boundary.

> the input, retrieved context, conversation history, and system prompt are sent to the configured LLM (OpenAI or Anthropic)
> — https://www.lumay.ai/ai-platform/core-engine · 'How It Works - Request Flow', step 2 'LLM reasoning'

**Open models** · Partly — A locally hosted model can be configured, but the vendor never names an open-weight family anywhere I read; only OpenAI, Anthropic and Azure OpenAI are identified by name.

> You can also configure LuMay to use your own Azure OpenAI resource or a locally hosted model.
> — https://www.lumay.ai/ai-platform/security · Security FAQ, expanded answer to 'Which AI model providers are used, and do they see our data?'

**Model choice** · Yes — Provider choice is an explicit design property: agents, flows and connectors are meant to survive a swap between providers or any compatible API.

> The engine is provider-agnostic: you can swap between OpenAI, Anthropic, or any compatible API without changing the agent configuration, the conversation flow, or any downstream connector.
> — https://www.lumay.ai/ai-platform/core-engine · Overview section, second paragraph

## Where you use it

**In a browser** · Yes — Web and chat are first-class interface channels, and the voice product adds a self-service dashboard with a real-time view from the Growth plan upward.

> Voice, chat, web, email and APIs.
> — https://www.lumay.ai/platform · Platform Architecture section, layer 01 'Interface'

## Whose machine it runs on

**Self-hosted** · Yes — Four postures are sold, up to on-prem Kubernetes and a fully air-gapped build, and the pricing page attaches published floors to each rather than leaving them to sales.

> Yes. LuMay offers SaaS, Bring Your Own Cloud, private cloud, and on-premises deployment models. BYOC and on-premises give full control over data residency, network boundaries, and access policies.
> — https://www.lumay.ai/ai-platform/security · Security FAQ, expanded answer to 'Can we self-host LuMay for complete data control?'

_Not established: Open source_

## What it costs to run

**How it meters** · Yes — A per-agent monthly subscription carries an included allowance, then meters: per-minute voice overage and per-active-minute connector charges stacked on the base fee, with carrier fees passed through.

> All connector charges are billed per active minute, only when the connector is engaged on a live call. No hidden fees and no separate per-query RAG charges.
> — https://www.lumay.ai/ai-products/voice-agent/pricing · 'Add-Ons and Charges - Transparent Add-On Pricing' introduction

**Free tier** · Partly — Only the voice line publishes a way in without a call, and it is a time-boxed trial rather than a standing free plan; the other eight agents start at $499 to $10,000 a month.

> All plans include a 14-day free trial. No credit card required.
> — https://www.lumay.ai/ai-products/voice-agent/pricing · 'Choose Your Plan' header, line directly beneath the Monthly/Annual billing toggle

**API access** · Yes — A REST API is checked on every voice plan including the $249 entry tier, with webhooks appearing at Growth and GraphQL plus private endpoints reserved for Enterprise.

> REST API
> — https://www.lumay.ai/ai-products/voice-agent/pricing · 'Choose Your Plan', Basic ($249/mo) card, checked feature bullet; the Feature Matrix marks REST API present on all four plans

**MCP server** · Partly — MCP is real and native but gated: five servers from the Growth plan and unlimited on Enterprise, while the Basic and Starter cards show it struck through.

> MCP (Model Context Protocol) servers let you extend Voxentis's AI with custom tools, data sources, and workflows - without writing a full integration. Available on Growth and above.
> — https://www.lumay.ai/ai-products/voice-agent/pricing · 'Your Questions, Answered' FAQ, expanded answer to 'What is an MCP Server?'

**Bring your own key** · Partly — Bringing a private or fine-tuned model is offered, but it sits in the custom-enterprise add-on table with no published rate, so the entry plans cannot buy it off the page.

> Deploy a fine-tuned or private LLM within Voxentis platform
> — https://www.lumay.ai/ai-products/voice-agent/pricing · 'Custom Enterprise' add-on table, 'Custom LLM / Bring Your Own Model' row, priced 'Custom quote'

## Buying it for a team

**A company can buy it** · Yes — An organisation can standardise several agents on one contract with one security review and one renewal, sold from a $150K/yr enterprise platform floor.

> Standardize multiple agents on a single governed contract - one security review, one renewal, one price.
> — https://www.lumay.ai/pricing · Section 05 - Enterprise Platform, introductory line above the LuMay Enterprise Platform card

**Seat model** · Partly — Seats are only one dimension among several and only on some agents: Insight and Legal price partly on users, while voice prices on agents and included minutes with no per-seat rate published.

> Data sources + users + queries
> — https://www.lumay.ai/pricing · Section 02 - The Pricing Model, agent rate table, Insight Agents row, 'Priced On' column

**Admin controls** · Yes — Constraints are enumerated concretely: least-privilege roles, tool allowlists, approval gates and spending limits, plus token-budget and latency caps set per agent in the runtime.

> Role-based access, least-privilege permissions, tool allowlists, input/output validation, prompt-injection protection, approval gates, spending limits, and full execution logs
> — https://www.lumay.ai/ai-agent-development-company · 'Our AI Agent Development Services', item 11 'Security & Governance'

**Audit log** · Yes — Every request carries a correlation ID linking API call, database queries, connector dispatches and model calls into one traceable chain, and access logs are offered for audit review.

> X-Correlation-ID on every request with end-to-end OpenTelemetry tracing.
> — https://www.lumay.ai/ai-platform/security · Security Controls table, 'Audit trail' row

**Single sign-on** · Partly — The platform page presents Keycloak SSO with SAML/OIDC as a standard control, yet the voice product prices SSO/LDAP/AD integration as a custom-quote enterprise add-on, so it is not simply included.

> Enterprise identity and access management integration
> — https://www.lumay.ai/ai-products/voice-agent/pricing · 'Custom Enterprise' add-on table, 'SSO / LDAP / AD Integration' row, priced 'Custom quote'

_Not established: Pooled budget_

## What happens to your code

**Opt out of training** · Yes — Non-training rests on the hosted default routing to Azure OpenAI inside Microsoft's boundary rather than on a customer-facing toggle; self-hosting or your own Azure resource is the stronger route.

> In the default SaaS configuration, it connects to Azure OpenAI, which processes prompts within Microsoft's enterprise data boundary and does not use your data to train models.
> — https://www.lumay.ai/ai-platform/security · Security FAQ, expanded answer to 'Which AI model providers are used, and do they see our data?'

**Data residency** · Yes — Residency follows the deployment model rather than a region picker: LuMay's own Azure tenancy for SaaS, your Azure or AWS subscription for BYOC, your VPC, or your data centre.

> The same images run in every deployment model - LuMay-managed SaaS, your cloud subscription, your private VPC, or your on-prem Kubernetes cluster.
> — https://www.lumay.ai/ai-platform/deployment · Overview section, first paragraph; the Deployment Models table's 'Where data lives' column names each location

**Getting out** · Partly — Customers keep rights to their content and session data is not retained, but no export tool is published, fees are non-refundable by default and rollover of unused minutes is never addressed.

> All fees are non-refundable unless otherwise stated in your enterprise agreement.
> — https://www.lumay.ai/terms-of-service · Section 7) Fees And Payment

**Certifications** · Yes — Two certifications are asserted outright rather than as alignment, with GDPR readiness and HIPAA handled through a BAA on private-cloud or on-premises deployments.

> The platform is SOC 2 Type II and ISO/IEC 27001 certified, and supports GDPR compliance-readiness for regulated environments.
> — https://www.lumay.ai/ai-products/voice-agent/pricing · 'Your Questions, Answered' FAQ, expanded answer to 'Is Voxentis built for HIPAA-regulated workflows?'

## Provenance

Every fact above comes from the vendor. Nothing here is independently corroborated yet.

## Coverage elsewhere

None recorded. Every fact above comes from the vendor and is not independently corroborated yet.

## What changed

- 2026-09-09 — Capabilities · Commercial terms · Free tier · Price

## Correcting this record

Anything marked "Not established" is a question we have not answered — it is not a No. If you represent LuMay, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth.

Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.