Lyzr Studio · Build Entry
Build agents here.
Run them in your cloud.
Lyzr Studio is one product delivered in three plans that grow with you — deployed inside your own VPC, packaged around where you are in your agent journey, not just the size of your business.
The idea
One platform, three plans
Every plan is the same Lyzr Studio, deployed into your own cloud environment, with a curated set of capabilities matched to your stage. As you add agents and volume, the platform scales with you — and so does the plan. You start where you are and move up when the work demands it, without re-platforming.
Pricing is simple and consumption-based: you prepay into a pool of platform credits, and usage draws them down. There are no seat licenses to count and no forced tier jumps — top up whenever you need more.
- Visual Agent Studio + SuperFlow workflows
- Knowledge Base retrieval & 100+ blueprints
- Voice agents with phone calling
- Core safety: prompt-injection, PII redaction, toxicity, versioning
- Runs inside your VPC
- Launchpad guided onboarding
- Everything in Lite, plus:
- Cognis cross-session memory & knowledge graphs
- Computer Agent, Text-to-SQL, Live Sources
- Full Govern layer + Agent Simulation Engine
- SSO, audit log, deeper analytics
- Value Sprint + dedicated success engineer
- Everything in Scale, plus:
- Agent Improvement Engine (AIE) — full
- ShadowLM model distillation
- Sub-account workspaces & custom roles
- On-prem / air-gapped deployment options
- 2 dedicated forward-deployed engineers, year 1
Which plan fits you
Matched to your agent journey
Studio Lite
You want to build agents the right way and get your first one into production without breaking anything. The focus is on building fast, with the safety essentials that make shipping to production safe. Success looks like a first production agent in weeks, visibly working.
Studio Scale
You have agents in production and you're scaling. Now you need full pre-production simulation, SSO, richer monitoring, and the beginnings of an improvement loop — catching drift, evaluating quality, and hardening what's live.
Studio Enterprise
You're running a fleet and thinking about cost per agent, drift across dozens or hundreds of agents, and governance across your whole estate. You need the full improvement and distillation loop, the deepest governance, and a path to govern externally-built agents too.
How pricing works
Tokenomics + credits
Credits. Whatever you prepay converts 1:1 into platform credits, burned down at $5 per million tokens as you use the platform.
Top-ups. Recharge in the same slice size, any number of times — $50k + $50k and so on. You never have to jump to the next tier just to keep going.
What "processed tokens" means
The meter counts the total tokens that flow through a run from end to end — everything that goes in and everything that comes out. A single run often coordinates more than one agent: a SuperFlow, for example, may have two or three agents working inside it, and it is metered once, across all the tokens that pass through the whole run, rather than separately per agent. That total also includes the work the platform does around your agents — orchestration, guardrail checks, memory operations, and evaluation or simulation cycles.
Where this rate applies
The $5 per million tokens rate applies to on-premise / VPC deployments. Lyzr Studio on SaaS and Architect are priced separately — we'll share that pricing on request.
What's included
Capabilities by plan
| Capability | Lite | Scale | Enterprise |
|---|---|---|---|
| Agent Studio (visual builder) | ✓ | ✓ | ✓ |
| Single-agent & manager orchestration | ✓ | ✓ | ✓ |
| SuperFlow (durable workflows) | ✓ | ✓ | ✓ |
| Architect (prompt-to-app) | ✓SaaS-connected | ✓SaaS-connected | ✓on-prem |
| GitAgent | ✓optional | ✓ | ✓ |
| Computer Agent | ✓optional | ✓ | ✓ |
| Lyzr Memory (basic session) | ✓ | ✓ | ✓ |
| Cognis Memory (cross-session + temporal) | — | ✓ | ✓ |
| Knowledge Base retrieval | ✓ | ✓ | ✓ |
| Knowledge Graph retrieval | — | ✓ | ✓ |
| 100+ agent blueprints | ✓ | ✓ | ✓ |
| Tools framework | ✓prebuilt + self-serve custom | ✓prebuilt + self-serve custom | ✓+ 5 custom integrations by Lyzr |
| MCP & A2A connectors | ✓ | ✓ | ✓ |
| Proxy Agent (wrap one external agent) | — | ✓ | ✓ |
| Voice agents + phone calling | ✓ | ✓ | ✓ |
| Data Query / Text-to-SQL | — | ✓ | ✓ |
| Model layer / BYO model & key | ✓ | ✓ | ✓ |
| Self-hosted open-weights model | — | ✓ | ✓ |
| Live Sources (SharePoint sync, etc.) | — | ✓ | ✓ |
| Capability | Lite | Scale | Enterprise |
|---|---|---|---|
| Prompt Injection Manager | ✓ | ✓ | ✓ |
| PII Redaction | ✓ | ✓ | ✓ |
| Toxicity Controller | ✓ | ✓ | ✓ |
| Secrets Detection | — | ✓ | ✓ |
| Allowed / Banned topics, Blocked keywords | — | ✓ | ✓ |
| Fairness & Bias Manager | — | ✓ | ✓ |
| Hallucination Manager | ◐groundedness | ✓+ Reflection | ✓+ LLM-as-Judge |
| Agent Simulation Engine | — | ✓10k+ scenarios | ✓10k+ scenarios |
| Audit Log | — | ✓ | ✓ |
| Traces & Monitoring | ✓basic dashboard | ✓deeper analytics | ✓full + custom |
| Agent Versioning (snapshot / rollback) | ✓ | ✓ | ✓ |
| RBAC | ◐standard roles | ✓custom roles | ✓custom roles |
| Team Management (unlimited seats) | ✓ | ✓ | ✓ |
| Sub-Account Workspaces | — | — | ✓ |
| Identity Broker / SSO | ◐basic auth | ✓OIDC SSO | ✓IdP-agnostic, SAML + OIDC |
| VPC-isolated deployment | ✓ | ✓ | ✓ |
| Compliance inheritance | ✓ | ✓ | ✓ |
| Capability | Lite | Scale | Enterprise |
|---|---|---|---|
| Agent Improvement Engine (AIE) | — | — | ✓full, all schedules & metrics |
| ShadowLM (model distillation) | — | — | ✓early access |
| Control Plane — cross-framework CI/CD | — | — | ＋optional add-on |
| Deployment model | Customer VPCSaaS optional | Customer VPCSaaS optional | VPC / on-prem / air-gappedSaaS optional |
✓ included◐ scoped / limited＋ paid add-on— not included
Getting to value
Onboarding & success support
Every plan comes with a support model designed to get you to a working production agent quickly — and to keep you moving as you scale.
Launchpad
A 30-day guided path to your first production agent, a blueprint and template library to build from, weekly group office hours, and async expert support. An optional hands-on Jump Start sprint is available if you'd like more direct help.
Value Sprint
A focused 4–6 week sprint with a Lyzr engineer to land your first two to three agents in production, followed by a dedicated success engineer across your account. Additional expert weeks are available whenever you need them.
Embedded FDEs
Two dedicated forward-deployed engineers embedded on your account through year one, working alongside your team to get agents into production and scale the fleet — backed by an ongoing success partnership and quarterly value reviews.
A note on the Control Plane
These three plans are the Build entry: the full Studio ships into your VPC, you build agents inside it, and it scales as you add volume — "build it here, run it in your cloud."
If you've already built agents externally on third-party frameworks (LangGraph, CrewAI, Strands, or custom) and your challenge is getting those agents to production and governing them centrally, that's the Control Plane — a separate motion priced flat plus per-node consumption. Enterprise customers can attach it as a bridge. We're happy to scope this alongside your Studio plan.
Good to know
Notes
- All figures are in USD. Token envelopes are planning estimates derived directly from the $5 / million platform rate and your prepaid amount; actual consumption depends on your configuration and usage.
- The platform charge is separate from LLM inference cost. You bring your own model key, or route through the Lyzr gateway at pass-through cost.
- Credit validity and top-up terms are confirmed in your order form.
- Compliance inheritance covers ISO/IEC 27001:2022 (certified), SOC 2 Type II (available under NDA), and a Satisfactory GDPR assessment.
- ShadowLM is in early access; availability is confirmed as part of your engagement.
- Individual capabilities and their plan placement may be tailored to your requirements during scoping.