Skip to Content
Pricing — Hire AI agents your team can trust

Pricing

Start free. Hire your first AI agent in under 5 minutes. No credit card required.

Curate-Me never marks up LLM API costs. You pay providers directly at their published rates. We charge for the governance platform — cost caps, PII scanning, approvals, observability, and managed runner access only when the private beta is enabled for your org.

What you get

Your AI assistant is a real workplace identity — a Microsoft 365 user with email, Teams presence, calendar, and a directory listing. Not a bot. On Starter and above, the assistant lives at curate@yourorg.curate-me.ai with up to 12 capability aliases (dev-team@, cfo@, security@, etc.) for per-task voice.

The platform has 22 Autopilot templates out of the box — code review, daily cost reports, security audits, customer support triage, sales research, and more. Connect a tool (GitHub, Slack, Linear, your data) and start the work from the dashboard or API. Browse all templates →


Plans

Free — Kick the tires

$0/month. No credit card.

  • 1 agent running at a time
  • 1,000 gateway requests/day for direct LLM access
  • Cost caps + audit trail — no surprise bills, ever
  • Community support

Best for: solo devs testing the platform, evaluating before adopting.

Sign up free → 


Starter — Hire your first AI agent

$29/month. 14-day free trial at the Pro tier on signup.

Everything in Free, plus:

  • 3 agents — pick from dev_team, cfo, security_audit, customer_support, sales_research, news_digest, or any of the 22 templates
  • 300K gateway requests/month
  • HITL approvals — high-cost or sensitive requests held for human review
  • Managed runners are not included by default in public self-serve Pro
  • Priority email support

Best for: small teams running one or two AI agents in production.


Growth — Hire a small team

$149/month. Most popular.

Everything in Starter, plus:

  • 5 agents running in parallel
  • 3M gateway requests/month at higher RPM
  • PII scanning + Presidio NER — production-grade sensitive-data protection
  • Runner beta eligible — managed runners and desktop streaming can be enabled for beta-approved orgs
  • Audit export (JSON) for compliance reporting
  • 99.9% uptime SLA

Best for: teams running multiple AI agents with cross-tool integration.


Enterprise — Run AI org-wide

Custom pricing. Built for regulated industries.

Everything in Growth, plus:

  • Unlimited agents and parallel runs
  • Bring your own VMs (BYOVM) (private beta) — run on your infrastructure
  • SSO / SAML + IP allowlisting
  • Custom HITL approval rules
  • SOC 2 audit export (JSON + PDF)
  • 1-year+ data retention
  • Dedicated engineer + 99.95%+ SLA

Best for: financial services, healthcare, defense, and any team where compliance is the buying criteria.

Contact sales →


vs. the alternatives

The “stitched stack” (Portkey + E2B + Langfuse) is fragmenting fast. Curate-Me is the integrated alternative that’s still independent.

PlatformStatus (May 2026)What you get
PortkeyAcquired by Palo Alto Networks (~$130M, Apr 2026)AI gateway, now powers Prisma AIRS
HeliconeAcquired by Mintlify (Mar 2026)LLM observability, in maintenance mode
LangfuseAcquired by ClickHouse ($15B Series D, Jan 2026)LLM observability + evals, ClickHouse-locked
LiteLLMIndependent OSSSelf-hosted proxy only — you run, monitor, scale, secure
Curate-MeIndependent, shipping weeklyGateway + governance + managed runners (private beta) + agent marketplace

See full comparisons →


Full feature matrix

For the technically inclined:

FreeStarterGrowthEnterprise
Price$0/mo$29/mo$149/moCustom
Your AI Assistant identityBot only (notifications@curate-me.ai)Real M365 user — curate@yourorg.curate-me.ai + 12 capability aliasesSame + up to 3 promoted specialists (dedicated cfo@, security@, etc. licensed identities)BYO Microsoft 365 — curate@yourdomain.com in your own tenant
Teams presenceBot framework only✓ Real user in directory, DM, calendar✓ + per-specialist presence✓ In your tenant
Capability slots1All 22 templatesAll 22 + concurrentAll 22 + concurrent + custom
Gateway requests1K/day300K/mo3M/moUnlimited
Managed runner / BYOVM beta accessEligiblePrivate rollout
RPM (requests per minute)1060300Custom
Data retention7 days30 days90 days1 yr+
Governance chain (15 stages)FullFullFullFull + custom
Cost caps + budgets✓ + hierarchical
PII scanningRegex (33 patterns)RegexRegex + Presidio NER+ custom rules
Content safety scanning
HITL approvals✓ + custom rules
Desktop streaming (VNC/RDP) (private beta)
MCP server (50 tools)
A2A agent communication
Skill scanner10/day10/dayUnlimited+ custom
Audit exportJSONJSON + PDF + SOC 2
Your machines (BYOVM) (private beta)
SSO / SAML
IP allowlisting
SupportCommunityPriority emailPriority + SLADedicated engineer
Uptime SLABest effortBest effort99.9%99.95%+

Annual billing saves ~20% on Starter and Growth plans.


Overage pricing

If you exceed your plan’s monthly allowance, overages bill at:

ResourceRate
Gateway requests$0.10 per 1,000 requests
Compute (lightweight runner)$0.10/hr
Compute (standard runner)$0.25/hr
Compute (heavy runner)$0.50/hr
Additional agent slot$25/mo per agent

Hard caps in the dashboard prevent unexpected charges. Set a monthly cap, and we’ll block traffic the moment you hit it — no surprise invoices.


FAQ

Can I switch plans? Yes. Upgrade immediately, downgrade at next billing cycle.

Free trial? All paid plans include a 14-day trial at the selected paid tier. No credit card required to start.

What counts as a gateway request? Each LLM API call through the gateway (completions, embeddings, image generation). Health checks and metadata queries are free.

Can I add extra agents without upgrading? Yes, $25/month per additional agent slot on any paid plan.

What if I want to migrate from Portkey, Helicone, or Langfuse? We’ve written migration guides for each — typically a 1-day lift since most apps just need an environment variable swap. See Curate-Me vs Portkey, vs Helicone, vs Langfuse.

Do you mark up LLM costs? No. You pay your provider directly at their published rates. Our pricing is for the governance platform only.

Is my data used to train models? No. We don’t train on customer data. Audit logs are stored in your own MongoDB/Redis (or ours, encrypted at rest).


Get started