Free
$0 / month
For developers and testing
Base includes
All plans include every validator and template. Compare monthly requests, rate limits, and analytics below — Developer includes a free trial (7 days); Growth and Production bill from day one.
Free includes 10,000 requests per month (hard cap), shadow mode on every tier, and 60 requests per minute?. Paid plans can opt into metered overage (Overage starts from $3 per additional 100,000 requests beyond your monthly allowance (depending on plan)). Developer adds envelope-encrypted Provider vault; Growth+ adds custom endpoints for locally hosted stacks (vLLM, Ollama, LiteLLM) or private cloud models without losing validation layer features.
Free is the foundation. Each upgrade keeps what you already have and adds higher limits or new capabilities — all validators and templates stay unlocked on every tier.
$0 / month
For developers and testing
Base includes
$69 / month
For growing developers, small teams, and production MVPs
Everything in Free, plus
$149 / month
Most popular. The sweet spot for teams building real production agents
Everything in Developer, plus
$399 / month
For companies that need reliable, high-volume AI in production
Everything in Growth, plus
Starting from $1,499 / month
For large organizations where AI failure is not an option
Everything in Production, plus
| Feature | FreeBase tier | DeveloperEverything in Free, plus | GrowthEverything in Developer, plus | ProductionEverything in Growth, plus | EnterpriseEverything in Production, plus |
|---|---|---|---|---|---|
| Usage limits | |||||
| Requests / month1 request = one full single-agent flow (input checks, output checks, and send/receive processing). In playbooks, each live agent step with an upstream call counts as 1 request (validations included). | 10,000 | 100,00010× Free | 1,000,00010× Developer | 5,000,0005× Growth | 10M–500M+High volume |
| Token usage trackingPrompt + completion totals are tracked in the dashboard for all plans. Hard caps (HTTP 402 token_cap_exceeded) apply only when you set an org or per-agent monthly token limit. | ✓ | ✓ | ✓ | ✓ | ✓ |
| Requests / minute | 60 | 3005× Free | 1,2004× Developer | 5,000 | 50,000+10×+ Production |
| Vanguard Orchestration LayerMulti-agent playbooks on every plan — the same deterministic layer that validates each step. Editor Simulate is free; each live agent step counts as one request toward your monthly quota. | ✓ | ✓ | ✓ | ✓ | ✓ |
| Max agents / flowMaximum number of agent-type steps in a single playbook | 2 | ↑ 5 | ↑ Unlimited | Unlimited | Unlimited |
| Max providers / flowMaximum distinct LLM providers across agent steps in one playbook | 1 | ↑ 2 | ↑ Unlimited | Unlimited | Unlimited |
| Pending human approvalsMaximum concurrent playbook runs waiting on a human approval decision | 3 | ↑ 20 | ↑ 100 | 100 | 100 |
| Log retention | 7 days | 7 days | 30 days | 90 days | 365 days |
| Overage (per 100k requests)Metered beyond monthly allowance on paid plans with an active subscription and overage enabled. Free is hard-capped. | Hard cap | $7 | $5 | $3 | Custom |
| Alerts & support | |||||
| SLAUptime commitment on paid production tiers | — | — | — | 99.5% | Up to 99.99% |
Full feature matrix with every capability is available on larger screens.
| Feature | FreeBase tier | DeveloperEverything in Free, plus | GrowthEverything in Developer, plus | ProductionEverything in Growth, plus | EnterpriseEverything in Production, plus |
|---|---|---|---|---|---|
| Usage limits | |||||
| Requests / month1 request = one full single-agent flow (input checks, output checks, and send/receive processing). In playbooks, each live agent step with an upstream call counts as 1 request (validations included). | 10,000 | 100,00010× Free | 1,000,00010× Developer | 5,000,0005× Growth | 10M–500M+High volume |
| Token usage trackingPrompt + completion totals are tracked in the dashboard for all plans. Hard caps (HTTP 402 token_cap_exceeded) apply only when you set an org or per-agent monthly token limit. | ✓ | ✓ | ✓ | ✓ | ✓ |
| Requests / minute | 60 | 3005× Free | 1,2004× Developer | 5,000 | 50,000+10×+ Production |
| Vanguard Orchestration LayerMulti-agent playbooks on every plan — the same deterministic layer that validates each step. Editor Simulate is free; each live agent step counts as one request toward your monthly quota. | ✓ | ✓ | ✓ | ✓ | ✓ |
| Max agents / flowMaximum number of agent-type steps in a single playbook | 2 | ↑ 5 | ↑ Unlimited | Unlimited | Unlimited |
| Max providers / flowMaximum distinct LLM providers across agent steps in one playbook | 1 | ↑ 2 | ↑ Unlimited | Unlimited | Unlimited |
| Pending human approvalsMaximum concurrent playbook runs waiting on a human approval decision | 3 | ↑ 20 | ↑ 100 | 100 | 100 |
| Log retention | 7 days | 7 days | 30 days | 90 days | 365 days |
| Overage (per 100k requests)Metered beyond monthly allowance on paid plans with an active subscription and overage enabled. Free is hard-capped. | Hard cap | $7 | $5 | $3 | Custom |
| Platform & validation | |||||
| All rules & templatesEvery tier — no pack gating by default | ✓ | ✓ | ✓ | ✓ | ✓ |
| All providers | ✓ | ✓ | ✓ | ✓ | ✓ |
| Basic dashboard | ✓ | ✓ | ✓ | ✓ | ✓ |
| Shadow modeAll tiers can run validators in shadow mode — observe failures and would-block metrics without blocking production traffic. Disable shadow mode per agent when you are ready to enforce. | ✓ | ✓ | ✓ | ✓ | ✓ |
| Response RepairDeterministic self-healing of model output before validators run. Available on every tier, including Free. | ✓ | ✓ | ✓ | ✓ | ✓ |
| Validation retries (re-ask)All tiers can block on validation failure (on_fail: block). Developer+ adds automatic re-ask retries when a validator is configured with on_fail: reask. | — | ✓ | ✓ | ✓ | ✓ |
| Analytics + request tracing | — | — | ✓ | ✓ | ✓ |
| Compliance workflowsPII, Safety, JSON, RAG — validators on all plans; tracing on Growth+ | Validators included | Validators included | ✓ | ✓ | ✓ |
| Provider VaultEncrypted storage for upstream API keys. Developer supports up to 2 providers per playbook (e.g. cheap + strong models); Growth+ is unlimited. Custom endpoints remain Growth+. | — | ✓ | ✓ | ✓ | ✓ |
| RBAC | — | — | — | ✓ | ✓ |
| Alerts & support | |||||
| Email alerts | — | ✓ | ✓ | ✓ | ✓ |
| Budget alerts | — | Basic | Full | Full | Full |
| Priority email support | — | Email alerts only | ✓ | ✓ | ✓ |
| SLAUptime commitment on paid production tiers | — | — | — | 99.5% | Up to 99.99% |
| Slack support | — | — | — | ✓ | ✓ |
| Enterprise controls | |||||
| Private Cloud / VPC | — | — | — | — | ✓ |
| Self-hosted deploymentSame web + engine + Postgres on your VPC or on-prem. Enterprise-only.Trust & deployment | — | — | — | — | ✓ |
| Custom rules & agent packs | — | — | — | — | ✓ |
| Dedicated Slack / Teams | — | — | — | — | ✓ |
| SLO p99 < 15 ms | — | — | — | — | ✓ |
| Audit logs | — | — | — | — | ✓ |
| BYOK / BYOVPC / BYOProvider | — | — | — | — | ✓ |
Stack annual billing with enterprise pilots where available. Enterprise teams negotiate committed-use rates.
Pay yearly on Developer, Growth, or Production and get two months free versus monthly. Toggle annual pricing above to see per-year totals.
Developer $690/yr · Growth $1,490/yr · Production $3,990/yr
Test with real production traffic. Need to verify latency, throughput, or compliance under real load? Get a dedicated, 14-day Sandbox environment with Production-tier limits and shared Slack engineering support.
No credit card required for qualified engineering teams — Contact Sales.
Paid plans with an active subscription can opt into metered overage beyond the monthly allowance. Free stays hard-capped until you upgrade.
Overage starts from $3 per additional 100,000 requests beyond your monthly allowance (depending on plan) — Developer $7, Growth $5, Production $3 / 100k when overage is enabled; Enterprise uses committed volume.
High-volume or multi-region deployments get negotiated rates, pooled limits, self-hosted deployment (VPC or on-prem), and custom SLA.
Trust and deployment · Contact sales for committed-use pricing
More billing questions? See the FAQ or contact us.
high-volume tier (10m–500m+ requests), self-hosted deployment (VPC or on-prem), custom rules and agent packs, dedicated Slack or Teams, and negotiated SLA with SLO p99 under 15 ms. Trust and deployment · Self-host docs
Contact salesPick a self-serve tier and tune monthly requests. Allowances: 10k Free, 100k Developer, 1M Growth, 5M Production — plus metered overage in 100k blocks (Overage starts from $3 per additional 100,000 requests beyond your monthly allowance (depending on plan)). For 10M+ volume, contact Enterprise sales.
Within your Growth allowance (1,000,000 included).
Estimated number of requests you expect across all agents each month before overage. 1 request = 1 full single-agent flow: input validation, output validation, and request/response processing. In playbooks, each live agent step with an upstream call counts as 1 request (validations included). Payload size and token count do not change request count.
Formula: $124.17 base plan + $0 overage (within included allowance) = $124.17
Subscription is billed yearly as $1,490 upfront. Estimated overage stays monthly on top of allowance.
About these numbers. This estimator uses public list prices with annual prepayment shown as a monthly amortization plus rounded metered blocks; actual invoices may add taxes, use different rounding, or include negotiated discounts or promotions. Monthly request volume here is an assumption — talk to sales for quotes beyond self-serve tiers.
$690/yr
2 months free vs monthly
7-day free trial
For growing developers, small teams, and production MVPs.
$1,490/yr
2 months free vs monthly
Most popular. The sweet spot for teams building real production agents.
$3,990/yr
2 months free vs monthly
For companies that need reliable, high-volume AI in production.
Starting from $1,499
For large organizations where AI failure is not an option.