Lyzr

Included quota

Community (free): 500 credits, 1 builder license, base models, 100MB knowledge base, 7-day logs, unlimited agents + users. Starter ($19/mo): ~2,000 credits/mo, 1 license, 7-day logs, more KB. Pro ($99/mo, $79 annual): 10,000 credits/mo, 1 license, all standard/leading models, a few Super Agents, 1GB KB, 3-month logs. Enterprise Cloud (custom): unlimited credits, 50+ builder licenses, premium/custom models, 120GB KB, 1-year logs, 24/7 support, 10 Super Agents, 48-hr integration SLA, Agent Entitlement Policy, Org General Intelligence, Lyzr Build Services (5 agents). On-Premise (custom): unlimited credits, custom licenses, logs retained forever, 20 Super Agents. (Some listings also show mid Teams $999 / Organization $2,499 tiers.)

What is public

Lyzr (lyzr.ai - enterprise agentic-AI platform; Agent Studio no-code/low-code builder on its 'safe & responsible AI' Agent Framework) publishes transparent low-end tiers, notable in a quote-heavy category: Community free ($0, 500 credits, 1 license, 100MB KB, 7-day logs), Starter $19/mo (2,000 credits), Pro $99/mo ($79/mo annual; 10,000 credits/mo, super agents, 1GB KB, 3-month logs), and custom Enterprise/On-Prem (unlimited credits, 50+ licenses, 24/7 support, on-prem/VPC). CRITICALLY, production is billed SEPARATELY per agent run - $0.08/run on Lyzr Cloud, $0.03/run on VPC/on-prem - plus LLM token pass-through; the subscription is a small part of the real bill.

Billing mechanics

Three cost layers: (1) a flat SUBSCRIPTION tier with a bundled monthly CREDIT allowance (500 Community / 2,000 Starter / 10,000 Pro / unlimited Enterprise) - credits meter build-time + premium operations (knowledge-base ops, tool calls, specific features); (2) PER-AGENT-RUN production charges billed on top - $0.08/run on Lyzr Cloud, $0.03/run on Lyzr VPC/on-prem; (3) LLM TOKEN costs passed through at provider rates (OpenAI/Anthropic/Google/Bedrock), accruing directly to the model provider. Annual billing gives two months free (Pro $99 to $79/mo). Extra credits top up one-time at $0.01/credit ($10/1,000 up to $5,000/500,000).

Variable cost rationale

High and multi-layered - flat subscription is minor; the real variable cost is per-agent-run production billing ($0.08 Cloud / $0.03 VPC-on-prem) scaling with invocation volume, PLUS LLM token pass-through (often the largest line), PLUS credit consumption for KB/tool/premium ops; deployment choice (Cloud vs VPC/on-prem) materially changes per-run economics

Additional watchouts

Runtime (per-agent-run) + LLM pass-through are the real bill - the cheap $19/$99 sub is a 'rounding error' at production scale (budget 5-10x); documentation gaps cause trial-and-error that wastes credits (top complaint); setup can feel heavy/complex for small teams; aggregator pricing is a mess (verify on official docs); thin public review/Reddit presence

Overage / add-ons

When monthly credits run dry, buy one-time top-ups at a flat $0.01/credit ($10 for 1,000, up to $5,000 for 500,000). Production agent runs are billed as you go ($0.08 Cloud / $0.03 VPC-on-prem per run) - there's no cap; LLM token usage bills directly from your model provider. Higher tiers raise credit bundles and lower effective run economics (VPC/on-prem at $0.03 vs Cloud $0.08).

Sales call required

Mixed (some tiers require a call)

Free / trial

Free

Lowest paid plan

$19/mo

Commercial notes

Positions Agent Studio as the first agent framework with safe/responsible AI natively in the core; strong in regulated verticals (BFSI, insurance, healthcare, IT services); pre-built agentic blueprints + multi-agent orchestration; 75% of customers run 2+ agents; NTT Data case study cites 70,000+ hours saved. NY-based; raised an $8M Series A (Rocketship.vc + Accenture Ventures, 2025) then a $14.5M round led by Accenture (March 2026) at a ~$250M valuation; 400+ enterprise customers (Accenture, AWS, Hitachi Energy, Publicis, AirAsia). Competes with Relevance AI, CrewAI, StackAI, Agentforce; builds agents - NOT a contact-data provider (pair with an enrichment layer for B2B data).

Key ambiguities

Subscription tiers are clear (Community/$19/$99/Enterprise) but aggregators show wildly conflicting numbers ($19 vs $99 vs $999 Teams vs Rs 143,928 vs $2,499 Organization) due to currency/outdated listings; Pro credit allowance is cited as 10,000/mo (= 120k/yr); some sources list intermediate Teams ($999) / Organization ($2,499) tiers between Pro and Enterprise; what exactly consumes 'credits' vs triggers a billable 'agent run' isn't fully itemized

Cancellation / refund

Community free (no card); Starter/Pro self-serve monthly or annual (annual = two months free, Pro $99 to $79/mo); one-time credit top-ups ($0.01/credit, no subscription); Enterprise/On-Prem custom-contracted (50+ licenses, build services); BYOM and on-prem keep data/IP in your environment

Support SLA / resale

Email/help-desk + chat + knowledge-base support (lower tiers); Enterprise adds 24/7 support, a 48-hr custom-integration SLA, dedicated build services, SSO, RBAC, audit logging, Human-in-the-Loop; SOC2/GDPR/ISO 27001 compliant; model-agnostic (OpenAI/Anthropic/Google/Bedrock + BYOM); deploy on Lyzr Cloud, VPC, or on-prem; native 'safe & responsible AI' (Hallucination Manager, PII redaction); integrates with Slack/Salesforce/Snowflake/Pinecone/Perplexity/etc.

Missing data

Aggregators show conflicting numbers (currency/outdated); intermediate Teams ($999)/Organization ($2,499) tiers appear in some listings; exact credit-vs-run consumption rules aren't fully itemized. Seed 'Community free (500 credits); Starter $19/mo (2k); Pro $99/mo (10k; $79/mo AE); Enterprise custom' is accurate for the subscription tiers - but the seed OMITS the key cost driver: production is billed per agent run ($0.08 Cloud / $0.03 VPC-on-prem) on top, plus LLM token pass-through, plus $0.01/credit top-ups.