Any agent, our skills
ERPAI ships skills you can load into any harness — Claude Code, ChatGPT, or anything MCP-capable. The agent reads records, writes them, and runs workflows on your account; your provider bills the model tokens.
Credit model
One meter for every operation
Pricing for headless SaaS
One credit meters everything. A credit is $0.005 of platform usage, and it goes a long way: reads, writes, imports, workflows, exports — every operation draws a published number of credits, most of them bundled a hundred to one. Your plan includes a monthly allowance, and most teams never spend past it.
What one credit buys
1 credit = $0.005
- Record reads100
- Record writes20
- Imported / synced records1,000
- Workflow runs10
- Code-node compute60 CPU-sec
- Triggers, webhooks or integrated starts100
- File upload10 MB
- per credit
- $0.005
- free grant to start
- $10,000
- per seat, ever
- $0
- for failed requests
- 0 credits
Plans
Allowances that scale with usage
Simple Credit Pricing
Pay for what you use. No per-seat fees. Every app included on every tier.
Free
$0
2,000,000 credits once, never expires, no card
- Core app platform access
- Unlimited users & agents
- Bring any agent: Proto, or your own with a key
- No automatic paid overage
- 1 automation in parallel
- 15 min active runtime
- 10 req/s external API rate limit
- 1 GB data storage
- Community support
Starter
$20
per month, 5,000 credits included, then $0.005 each
- Everything in Free
- Allowance refreshes monthly
- 1 automation in parallel
- 15 min active runtime
- 25 req/s external API rate limit
- 10 GB data storage
- Email support
Team
$99
per month, 25,000 credits included, then $0.005 each
- Everything in Starter
- 2 automations in parallel
- 2 hr active runtime
- 100 req/s external API rate limit
- 100 GB data storage
- 99.9% uptime SLA
- Email + chat support
Business
$499
per month, 150,000 credits included, then $0.005 each
- Everything in Team
- 5 automations in parallel
- 8 hr active runtime
- 250 req/s external API rate limit
- 500 GB data storage
- SSO / SAML, 90-day audit log
- Priority support
Enterprise
Custom
Custom allowance, then $0.005 per overage credit
- ERPAI-Managed AI at provider cost × 1.087
- Custom allowance and invoicing
- Custom active runtime
- 99.99% uptime SLA
- 1,000 req/s external API rate limit
- Custom storage
- SSO / SAML / SCIM
- Invoicing, DPA, custom terms
- Dedicated TAM
- 24/7 P1 support (1-hour)
Agentic SaaS
Any harness, same permissions
Run it with any agent.
Every ERPAI app is an API surface, and ERPAI ships skills any agentic harness can load. Point an agent at your workspace and it reads records, writes them, and runs workflows — same permissions, same audit trail, same credit meter as a person.
Estimate
Model your monthly usage
What would your month cost?
List and get calls from dashboards, pages and the API
Priced per call, regardless of result size
Create, update, delete
Bulk loads and scheduled data syncs
Any automation execution, however many steps
Scheduled, webhook and integrated starts
Orchestration only — the model runs on your own account
Beyond the compute included in every run — most workflows never need this
On Starter — your usage fits inside the included 5,000-credit allowance.
Monthly bill
Base plan$20.00
Platform usage · 4,200 credits$21.00
Included allowance applied−$21.00
Total$20
Credits draw in the order that favors you: allowance first, then pay-as-you-go.
Rate card
Credits published by operation
Every operation, priced in the open
| Operation | Price | Per unit | What that means |
|---|---|---|---|
| Reading data | |||
| List, get or filter records | 1 credit | / 100 calls | $0.00005 a call — refresh a dashboard all day |
| Search and aggregate | 1 credit | / 20 calls | $0.00025 a call, whatever the result size |
| Export CSV or PDF | 2 credits | / export | $0.01 — real rendering work |
| Writing data | |||
| Create, update or delete | 1 credit | / 20 writes | $0.00025 a write |
| Bulk import or scheduled sync | 1 credit | / 1,000 records | $0.000005 a record |
| File upload | 1 credit | / 10 MB | Storage billed at $0.20/GB-month |
| Automation | |||
| Workflow run | 1 credit | / 10 runs | $0.0005 a run, including 5 CPU-seconds of code-node compute. Data steps bill as the reads and writes they make; AI steps bill as AI requests; branches and parked waits are free |
| Code-node compute beyond included | 1 credit | / 60 CPU-sec | $0.30 per CPU-hour of sandbox CPU — 99% of workflows never touch this |
| Trigger, webhook or integrated start | 1 credit | / 100 | $0.00005 per activation |
| Integration steps inside a runNetSuite, email, HTTP, Slack, Postgres… | 0 extra | — | no extra orchestration charge; reads and writes still use their published rates |
| Report steps in workflows | 0 extra | — | a report step is the reads it makes. Dashboards and reports inside the app aren’t a billable item at all |
| AI request orchestration | 1 credit | / 10 requests | $0.0005 — routing, context and logging around each AI call |
Guardrails
Caps, alerts, and ledger visibility
You should not be able to get a surprise bill
Caps are on by default
Paid plans start with a monthly spend cap. Reach it and billable operations pause — reads keep working, deletes stay available for cleanup — until you raise it or the month turns.
Alerts before it matters
80%, 90% and 100% of your cap or allowance, once per threshold per month, to email or a webhook you choose.
The meter is public
Every operation shows its credit draw in the ledger, live. Filter by app, user or operation, and export it all as CSV whenever you want.