Usage metering
Slab5 uses usage-based metering. Customers choose a plan with included monthly usage, then Slab5 records metered work against that usage balance so teams can understand cost without reasoning about every individual request price.
Usage-based metering
Slab5 meters individual operations, but the product presents usage primarily as dollar-denominated usage value:
- BYOK Developer is free, BYOK-only, and includes capped platform usage for evaluation.
- Builder includes
$99of monthly usage. - Team includes
$299of monthly usage. - Business / Agency includes
$799of monthly usage. - Prepaid usage balance can extend usage up to the plan cap.
- Enterprise usage is custom.
Per-operation rates still exist so usage is fair, auditable, and explainable.
What is metered
Slab5 meters work that creates platform cost, AI cost, execution cost, storage cost, transfer cost, or operational risk:
- REST API operations
- MCP tool calls
- Webhook delivery attempts
- Asset upload intents
- Storage and egress
- Analytics Sites event collection
- Object Analytics and BI refresh work
- Semantic Search, RAG in a box, and Vector Store operations
- Workspace Assistant usage
- Slab5 Data Agents usage
- AgentGrid-recorded AI, tool, and retrieval usage
- Execution jobs and scheduled operational work
CMS REST reads are metered API reads. Production websites should cache published CMS reads with build-time fetches, server-side revalidation, or a publish-to-cache flow instead of calling Slab5 CMS on every pageview.
Slab5 Guide and support-agent style help are included so teams can get answers and troubleshoot without a separate support meter.
The BYOK Developer plan is intended for developers evaluating Slab5 with their own OpenAI or Claude key. It includes one workspace, capped API/MCP usage, capped app events, capped business records, capped storage, capped AgentGrid runs, capped BI refreshes, and daily-or-slower recurring workflows. Slab5 managed AI is disabled on this plan.
How usage is counted
Usage events map to billing dimensions such as light REST reads, search REST reads, standard REST writes, critical REST writes, MCP read tools, MCP write tools, webhook delivery attempts, asset upload intents, storage GB-month, egress GB, app events, BI refreshes, and agent execution work.
Lightweight reads cost less than writes. Agent write tools cost more than direct API writes because they often represent higher-value workflow automation. Critical writes, such as invoices, payments, accounting, and archive/destructive actions, are separated for clearer cost and risk accounting.
Current base meters:
| Meter | Example price |
|---|---|
| Light REST read | $0.00005 |
| Search REST read | $0.00015 |
| Standard REST write | $0.001 |
| Critical REST write | $0.0025 |
| MCP read tool | $0.0005 |
| MCP write tool | $0.003 |
| Webhook delivery attempt | $0.00075 |
| Asset upload intent | $0.0015 |
| Storage | $0.15/GB-month |
| Egress | $0.15/GB |
| Slab5 Guide | $0 |
| Workspace Assistant safe summary | $0.015 |
| Data Agent reviewer | $0.10 |
| Light Data Agent execution job | $0.25 |
| Standard Data Agent execution job | $1.00 |
| Heavy Data Agent execution job | $5.00 |
| AgentGrid run logging | $0.005 |
| AgentGrid retrieval hit | $0.001 |
| Vector Store operations | Metered by operation, embedding tokens, and retrieval size |
| BI Intelligence query execution | Metered as Data Agent execution and BI Intelligence usage |
| Slab5 managed AI tokens | Provider-metered |
| BYOK AI tokens | Observed telemetry |
Failed authentication requests are not billed because billing checks run after credential validation.
Vector Store and intelligence usage
Slab5 meters semantic and retrieval features through the shared Vector Store usage model.
Vector Store usage can include:
- collection creation and archive events
- document upserts and deletes
- embedding input tokens
- vector upload bytes
- vector query events
- estimated query scan bytes
- vectors returned
- estimated bytes returned
Semantic Search, RAG in a box, BI Explore, AgentGrid memory, AgentGrid workflow discovery, and future AI-aware features can all use Slab5 managed retrieval. The service is shared operationally, but indexes and collections are scoped by tenant, workspace, and use case.
Workspace admins can open Usage → Intelligence usage to review:
- Vector Store summary usage
- RAG collection usage by collection
- Semantic Search usage
- BI Intelligence usage
- AgentGrid memory and workflow discovery usage
- embedding pricing by dimensions and event count
RAG in a box exposes collection-level drill-downs so admins can see which app, workflow, MCP client, or integration is driving usage.
When a Vector Store operation is performed through an app-associated API key, Slab5 records app attribution in usage metadata so admins can filter usage by app and inspect the app detail rollup.
Managed AI and BYOK usage
AgentGrid separates Slab5 managed AI usage from bring your own key usage.
| AI mode | Billing behavior | Usage visibility |
|---|---|---|
| Slab5 managed AI | Slab5 provides the model key and records managed AI token usage against the workspace usage balance. | Owners can inspect managed AI usage, budgets, run costs, and AgentGrid traces. |
| Bring your own key (BYOK) | The workspace provides the OpenAI or Claude key. Slab5 does not bill provider token charges; the provider bills the customer directly. | Slab5 records approximate BYOK token telemetry separately for visibility, debugging, and budget awareness. Platform meters still apply. |
The AgentGrid Settings mode switch is authoritative. Stored BYOK keys can exist without being used. They are used only when BYOK mode is selected and a BYOK provider is selected for the workspace or agent.
BYOK Developer workspaces must use BYOK mode. Slab5 still records approximate BYOK token telemetry for visibility, but the model provider bills the customer directly.
Storage and egress
Storage is metered from asset usage snapshots. Slab5 records stored asset bytes by workspace and provider, converts the daily snapshot into GB-month usage, and debits only unprocessed snapshots.
Egress is recorded as observed or estimated transfer telemetry. This gives workspace owners visibility into download and data-transfer usage while keeping billing explainable. Plans can include a reasonable transfer allowance; transfer above the included allowance is visible through usage events and can be applied to usage balance.
Usage visibility
Workspace owners and admins should be able to see:
- Current plan and included monthly usage
- Usage consumed in the current billing period
- Remaining prepaid usage balance
- Monthly cap and low-balance warnings
- Metered usage by dimension
- Slab5 managed AI usage separately from BYOK token telemetry
- Storage and egress usage
- Request IDs tied to API, MCP, webhook, storage, Data Agents, and AgentGrid activity
- Billing pause and recovery states
Usage events remain the operational source of truth for billing views, audit investigation, request correlation, support troubleshooting, and upgrade decisions.
App attribution
Usage can be grouped by registered app when an API key or operation is associated with an app.
The Usage screen supports:
- All apps
- No app
- First-party app
- Specific app
App-attributed usage is useful for:
- separating production app traffic from internal tools
- identifying which app created Vector/RAG spend
- comparing failed request rates by app
- reviewing BI Intelligence and Vector Store events by product boundary
- finding legacy keys that still show as no-app activity
Use the app detail view to see an app-level usage rollup, top API keys, recent usage events, recent audit events, and app-owned Vector/RAG collections.
Pause behavior
When a workspace exhausts available usage balance and cannot add more within its cap, billable operations pause. Slab5 keeps recovery paths available: login, billing visibility, usage visibility, credential review, credential revocation, and data-reduction workflows remain accessible.
Upgrading from BYOK Developer to Builder usually makes sense when a developer needs a production workspace, Slab5 managed AI, more schedules, higher workflow volume, or more storage. Upgrading from Builder to Team usually makes sense when a team needs more than 2 workspaces, more schedules, higher app-event volume, more workflow runs, more API/MCP usage, or longer retention.
Upgrading from Team to Business / Agency usually makes sense when a team manages multiple products, client workspaces, higher BI refresh volume, heavier AgentGrid automation, or higher support requirements.
Moving to Enterprise usually makes sense when usage needs custom limits, SSO, custom retention, security review, SLA terms, dedicated execution, audit exports, or custom deployment requirements.