free
$0
Free forever
All 5 agents, starter playbooks. Drafts only.
- 1,000 credits at signup, plus 100 every Monday after that
- Every connector live, Gmail and GitHub included
- Today panel + ⌘K, audit logs, community support
- No card required
Free with a thousand credits to start, plus a Monday topup that keeps you going. The same credit pattern Cursor, Replit, and Bolt use when you're ready to put the rest of the crew to work.
$0
Free forever
All 5 agents, starter playbooks. Drafts only.
$31/ month
Billed yearly · save $192/yr
Auto-approval, custom playbooks
$78/ month
Billed yearly · save $468/yr
Multiple apps, shared playbooks
$199/ month
Billed yearly · save $1176/yr
Priority queue, more concurrency. Best value.
Run out of monthly credits? We pause the crew. Buy a top-up pack, upgrade, or wait for the next cycle and carry on with what rolled over (paid tiers carry unused credits forward one month). Overage is opt-in, not a default, so your spend stops at your budget. No surprise four-figure bills, ever.
A credit is one unit of agent work. Every step the crew runs costs credits, priced from the work it actually does. Frankie classifying an email lands light. Hank reasoning over a churn signal lands heavier. The schedule below shows what each kind of step typically costs, and most land right around those numbers. It's the same credit pattern Cursor, Replit, Lovable, and Bolt use, and your monthly budget keeps the total predictable.
| step type | typical |
|---|---|
| Light step (classify, lookup, simple draft) | 1 |
| Memory operation (read or write to agent memory) | 1 |
| Schedule or dispatch action (cron, webhook fire) | 1 |
| MCP tool call (cached) | 1 |
| MCP tool call (uncached, large response) | 3 |
| Standard step, cached prompt | 4 |
| Standard step, uncached complex work | 8 |
| Deep reasoning step | 30 |
| Long-context analysis (scales by content length)scales by token count | 10–50 |
typical costs · actual credits track the work each step does · we publish any rate change in advance
A typical Lou run when Stripe fires a payment_failed webhook. Most land around 6 to 10 credits.
step 1: classify decline code1 creditsstep 2: look up customer history1 creditsstep 3: draft recovery email4 creditsstep 4: schedule follow-up1 creditstypical run costabout 7 creditscredit balanceabout 9,990 / 10,000
About what a run costs when the agent has to do real reasoning. Heavier runs land higher.
step 1: classify customer segment1 creditsstep 2: analyze churn signals8 creditsstep 3: reason about retention30 creditsstep 4: draft personalized offer4 creditsstep 5: validate against rules1 creditstypical run costabout 44 creditscredit balanceabout 9,955 / 10,000
Plan credits renew monthly. When a month runs hot, top-up packs add more without an upgrade.
| top-up pack | credits | price |
|---|---|---|
| Small | 5,000 | $27 |
| Medium | 15,000 | $67 |
| Large | 50,000 | $197 |
one-time purchase · stacks on your monthly balance · good for 12 months
Credit prices cover more than raw inference — durability, observability, integrations, audit logging, and the platform layer itself. If you'd rather pay direct API rates and skip the platform markup, BYOA is available on every paid tier.
Grouped by category. Pricing & credits, agents & playbooks, workspace & seats, limits, and trust. The full row breakdown for the founder who wants to see it before signing up.
| Feature | free | growth | growth+ | scale | enterprise |
|---|---|---|---|---|---|
| pricing & credits | |||||
| Annual price | $0 | $31 / mo | $78 / mo | $199 / mo | Custom |
| Monthly price | $0 | $47 / mo | $117 / mo | $297 / mo | Custom |
| Credits | 1,000 + 100/wk | 10,000 credits / mo | 30,000 credits / mo | 100,000 credits / mo | negotiated |
| Credit rollover | — | 1 month | 1 month | 1 month | Custom |
| agents & playbooks | |||||
| All five agents (Lou, Frankie, Hank, Dale, Inky) | ✓ | ✓ | ✓ | ✓ | ✓ |
| Drafts vs auto-send | Drafts only | Both | Both | Both | Both |
| Bounded customization (tone, thresholds, signatures) | ✓ | ✓ | ✓ | ✓ | ✓ |
| Custom playbook builder | — | Yes | Yes · cross-app sharing | Yes · cross-app sharing | Yes · cross-app sharing |
| BYOA — bring your own Claude / MCP | — | Yes | Yes | Yes (priority MCP) | Yes (dedicated MCP) |
| workspace & seats | |||||
| Users per workspace | 1 | 2 | 5 | 15 | Unlimited |
| Apps per workspace | 1 | 1 | 3 | 10 | Unlimited |
| Multiple workspaces | — | — | — | — | ✓ |
| limits & performance | |||||
| Run history retention | 30 days | 90 days | 1 year | 2 years | Custom |
| Concurrency | 1 | 3 | 8 | 25 | Custom |
| Priority execution | — | — | — | ✓ | ✓ |
| support & trust | |||||
| Support | Community | Priority email | Priority email | Dedicated CSM, Slack | |
| Founder office hours | — | — | — | ✓ | ✓ |
| SOC 2 | — | — | — | Readiness assessment | Full evidence package |
| Audit logs | 30 days | 90 days | 1 year | 2 years | Custom |
Scale customers can book a 30-minute slot one Friday afternoon a month. Topics welcome: product feedback, operational questions, strategy, just talking shop. The COO is primary host; CEO or CTO pinch-hits.
Subject to availability — not an SLA, not a contract commitment, just a real conversation we make time for. Past about thirty Scale customers we'll evolve the format (group sessions, quarterly roundtables) so we can keep doing it well.
Credits are how the agents account for work. Each step the crew runs, Frankie classifying an inbound, Lou drafting a recovery, Hank reasoning about churn, costs a small number of credits priced from the work it does. A typical recovery lands around 7 credits and a heavy reasoning play closer to 44, and both vary a little run to run. Same primitive Cursor, Replit, and Bolt use.
1,000 credits at signup plus a Monday topup that brings your balance back to 100 every week, indefinite. Drafts-only across all five agents, no auto-send, no BYOA. 100 credits a week buys roughly 14 typical recovery runs, 10–15 changelog updates, or 3 heavy reasoning plays. No card required.
When your monthly credit budget runs out, we pause the crew. From there you can buy a top-up pack (5,000 to 50,000 credits, good for 12 months), turn on auto-overage, or just wait for the next cycle and pick up with whatever rolled over (paid tiers carry unused credits one month). Because overage is opt-in, your spend stops at your budget until you choose otherwise. No surprise four-figure bills, ever.
Annual is 33% off across all paid tiers, Growth saves $192/year, Growth+ saves $468, Scale saves $1,176. Monthly works if you'd rather not commit a year up front. Annual customers also get priority on early access and beta features. Toggle at the top of this page.
First 14 days, full refund, no questions. After that, we'll prorate the unused portion of your annual plan if you cancel mid-term. Monthly plans cancel at the next billing cycle. Refunds go back to the original payment method within 5–10 business days.
Yes, anytime. Downgrades take effect at the next billing cycle so you keep what you paid for through the current period. Unused credits don't carry across tiers — they expire on downgrade — but your data stays. Historical Lou recoveries, Hank flags, and Inky entries remain intact. Re-upgrade whenever.
No. Your customer data, your Stripe records, your inbox — none of it goes into model training. We use Anthropic's Claude under their commercial terms with no-training-on-data guarantees. This is a contractual commitment in our Privacy Policy.
Bring Your Own Anthropic. On any paid tier, connect your own Claude or Claude-via-MCP account and pay Anthropic directly, so platform work doesn't draw down your monthly credits. Subscription price stays the same, and the wedge is for technical users who want direct API economics. Scale routes via priority MCP, and Enterprise gets dedicated infrastructure.
A workspace is your business — one Stripe subscription, one team, one billing relationship. An app is a single product within it. Most founders run one workspace with one app. Portfolio founders managing 2–3 products use Growth+'s 3-app slot under a single workspace, instead of paying for three separate Growth subscriptions.
Bounded customization (every tier from Free up) means adjusting the shipped playbooks: tone, thresholds, signatures, segments, send timing. Custom playbooks (Growth and up) let you author flow logic from scratch with version control and AI-assisted authoring. Cross-app sharing of custom playbooks starts at Growth+.
Scale customers can book a 30-minute slot one Friday afternoon a month — the COO primary, CEO or CTO as backup. Topics: product feedback, ops questions, strategy, just talking shop. Subject to availability, not an SLA — closer to a real conversation we make time for.
Custom integrations, SLA commitments, compliance packages (SOC 2 evidence, BAAs), or sustained usage above what Scale comfortably covers. Pricing scales with credit allocation. Email [email protected]. We'll come back with a scoped proposal within a week, no calls until you ask for one.
The crew is new. Reply to any of our welcome emails, or email us — one of the cofounders will write back.
Email us →