API pricing
Pay for what you call
The Prava AI API is billed on usage alone. You pay per 1,000 calls, or per 1M tokens for LLM endpoints, with lower unit prices as volume grows.
Billing
How API billing works
Every request is metered. You see what you use as you use it, and you are billed once a month for what was used.
Note. All prices on this page are indicative and subject to confirmation. They are shown as placeholders until SwiftSafe publishes final rates. Where you sign an Order Form, the prices and terms in that Order Form govern.
Metering
Per call or per token
Most endpoints are metered per 1,000 calls. Prava AI LLM endpoints are metered per 1M tokens, with input and output counted separately. A few families use a unit that fits the work, such as per sample or per asset.
Invoicing
Monthly in arrears
Usage is totalled at the end of each calendar month and invoiced after it. You pay for what was used, not for what was forecast.
Visibility
Usage in the dashboard
The dashboard shows usage by endpoint family, API key and day, with an estimate of the month to date.
Control
Spend caps and alerts
Set a monthly spend cap per account or per key, and alerts at thresholds you choose. When a hard cap is reached, further billable calls are refused until you raise it.
Scope
Separate from seat plans
API usage is billed on its own. It does not change your seat plan, and seat plans do not include metered API usage unless your Order Form says so. See plans and pricing.
Currency
USD or INR
Prices are set in US dollars. Customers billed through the Indian company may be invoiced in Indian rupees. See billing and tax below.
Rates
Price by endpoint family
One unit price per family, before volume discounts. All prices are indicative until confirmed.
| Endpoint family | Unit | Price | Notes |
|---|---|---|---|
| Detection and telemetry ingestion | Per GB, or per 1,000 events | [US$ x.xx per GB] / [US$ x.xx per 1,000 events] | One unit applies per account, agreed at onboarding. |
| Threat intelligence lookups | Per 1,000 calls | [US$ x.xx] | Indicator, actor and vulnerability lookups. |
| Dark-web exposure queries | Per 1,000 calls | [US$ x.xx] | Queries are limited to domains, brands and identities you are authorised to monitor. |
| Malware sandbox submissions | Per sample | [US$ x.xx] | Static and dynamic analysis of one file or URL. |
| Vulnerability and attack-surface scans | Per asset per month | [US$ x.xx] | An asset is a host, domain, IP address or application in scope during the month. Scan only what you are authorised to test. |
| Prava AI LLM, input | Per 1M tokens | [US$ x.xx] | Tokens sent in prompts and context. |
| Prava AI LLM, output | Per 1M tokens | [US$ x.xx] | Tokens generated in responses. |
| Compliance evidence API | Per 1,000 calls | [US$ x.xx] | Reading and submitting evidence for controls. See compliance for what the platform is designed to align with. |
| Webhooks | Per delivery | Included | Outbound event deliveries and retries are not billed. |
Volume
Volume tiers
The more you use in a month, the lower the unit price. Tiers are measured per endpoint family and reset each month.
| Tier | Monthly volume | Discount on list price |
|---|---|---|
| Tier 1 | Up to [Tier 1 ceiling] | List price |
| Tier 2 | [Tier 2 range] | [x% off] |
| Tier 3 | [Tier 3 range] | [x% off] |
| Tier 4 | Above [Tier 4 floor] | [x% off], or a committed rate by Order Form |
Discounts apply to the volume within each tier, not retrospectively to the whole month. Committed-use rates for steady, high volumes can be agreed in an Order Form.
Start
Free allowance and sandbox
Build and test before you pay.
Allowance
Monthly free usage
Each account receives [Free allowance per month] before metered billing begins. Unused allowance does not roll over.
Sandbox
A separate test environment
Sandbox keys return realistic responses against test data and are not billed, within [Sandbox limits]. Sandbox data is kept apart from production.
Limits
Rate limits by plan
Limits protect the service for every customer. Responses include headers showing your remaining quota.
| Plan | Requests per minute | LLM tokens per minute | Concurrent sandbox submissions |
|---|---|---|---|
| Pro | [Pro RPM] | [Pro TPM] | [Pro concurrency] |
| Max | [Max RPM] | [Max TPM] | [Max concurrency] |
| Ultra Max | Set in your Order Form | Set in your Order Form | Set in your Order Form |
Requests over the limit receive HTTP 429 with a retry hint. Limits can be raised on request where your usage pattern supports it.
Invoices
Billing entity, tax and invoices
Americas
SwiftSafe LLC
Customers in the Americas are billed in US dollars by SwiftSafe LLC, the United States office.
India and elsewhere
SwiftSafe Cybersecurity Technologies India Pvt Ltd
Other customers are billed by the Indian company, in Indian rupees or US dollars.
Tax
Prices exclude tax
Applicable taxes, such as GST or sales tax, are added to the invoice where required. Provide tax registration details in the dashboard so they appear on invoices.
Invoices are issued monthly, list usage by endpoint family and tier, and can be downloaded from the dashboard. Payment terms are [Payment terms] unless your Order Form says otherwise.
Use and data
Fair use and data handling
Fair use
Use it as intended
API use is subject to the Acceptable Use Policy. Scanning, testing and dark-web queries are limited to assets and identities you own or are authorised to assess. Traffic designed to evade rate limits or resell access may be suspended.
Data
Processed in your region
Requests are served from the regional endpoint for your account, and customer data is stored and processed only in data centres in your own region. How API data is used by Prava AI models is set out in the AI data use policy.
FAQ
Questions
Do I need a seat plan to use the API?
[Confirm whether a seat plan is required for API access]. API usage is billed separately from seats in either case.
Are failed requests billed?
Requests rejected for authentication or rate limits are not billed. The full rules for errors and empty results are [Billing rules for failed and empty requests], and your Order Form governs.
How are LLM tokens counted?
Input and output tokens are counted separately and billed per 1M tokens. The response includes token counts so you can reconcile them with your invoice.
Can I set a hard limit on spend?
Yes. Set a monthly cap per account or per API key in the dashboard. You can also set alerts at any threshold below the cap.
Will prices change?
The prices on this page are indicative until confirmed. Once published, any change to list prices will be announced in advance, with the notice period set out in the Service Terms. Rates in a signed Order Form apply for its term.
Is there an annual commitment?
No. Usage is billed monthly with no lock-in. Committed-use rates are available if you want them.
Next step
Estimate your usage with us
Tell us what you plan to call and how often. We will confirm rates in writing before anything is charged.