Question 1

What does one API call actually return?

Accepted Answer

One call returns a structured evidence card for a specific ingredient and outcome pair. Each card includes a GRADE certainty rating (A through D), dose-match verdict (MATCH, PARTIAL, or MISMATCH against your labelled dose), a list of PubMed IDs with study-level summaries, and an HMAC attestation that lets you prove the card is unaltered. The card is deterministic: the same ingredient and outcome return the same card against the same corpus version.

Question 2

What are the pricing tiers?

Accepted Answer

Free: 10 Attested Evidence Cards per month, no credit card required. Developer: $49/month, full corpus access, commercial-use rights, monthly delta feed, standard support. Growth: $299/month, embeddable Attested Evidence Card widget, attribution removed, delta webhooks on your ingredient stack. Scale: $799/month, white-label attribution, evidence monitoring alerts, priority corpus refresh SLA, dedicated onboarding. Enterprise: custom pricing, full white-label and platform embed license, private corpus option, negotiated SLA.

Question 3

What is the founding price offer?

Accepted Answer

Developer access is available at $49/month for early signups, locked for 24 months regardless of where public pricing goes. This is the founding price. In exchange for locking in early: full corpus access, commercial-use rights, a monthly delta feed so you know when evidence shifts, and a direct channel to the roadmap as the API surface evolves. The founding price is available to early signups only and is not available after the introductory period closes.

Question 4

How does HMAC attestation work?

Accepted Answer

Each evidence card is signed with an HMAC using a server-side key. The response includes the card payload plus the signature and issuer public key URL. You verify locally by recomputing the HMAC over the canonical payload and comparing it to the included signature. If they match, the card is byte-for-byte identical to what the server produced. Your users can verify independently that you have not altered the evidence.

Question 5

How do rate limits work?

Accepted Answer

Free tier: 10 cards per month. Developer plan: 1,000 cards per month with burst headroom. Growth: 10,000 cards per month. Enterprise: negotiated. Rate limit headers are included in every response. The free tier is designed for evaluation and small integrations. If you need higher limits during evaluation, contact us.

Question 6

Is there a quickstart?

Accepted Answer

Yes. The docs at probara.dev/docs/api walk through authentication, your first request, and how to parse the evidence card response. The demo at probara.dev/demo lets you run live queries against 8 pre-seeded ingredient and outcome pairs without an API key.

Question 7

How is the evidence corpus maintained?

Accepted Answer

The corpus is built from PubMed and MEDLINE clinical literature only. No grey literature, no press releases. Studies are indexed by PMID and versioned. The corpus version is included in every evidence card response so you know exactly which dataset version produced the card. Corpus updates are announced in the changelog.

Question 8

Does the API cover all supplements?

Accepted Answer

Coverage is deep on high-research-volume ingredients: adaptogens, common vitamins and minerals, amino acids, botanical extracts with substantial RCT literature. We index what PubMed has. If your use case requires an ingredient not yet covered, the Enterprise tier includes custom corpus additions.

Question 9

Can I use this for regulatory submissions?

Accepted Answer

The API returns research-tool output based on published biomedical literature. It is not legal or regulatory advice and does not constitute Competent and Reliable Scientific Evidence as defined by the FTC. That determination requires review by qualified regulatory professionals. What the API gives you is a structured, traceable, reproducible evidence summary that is a better starting point for that review than manual literature search.

Vendor	Purpose	Data Processed	Location
Cloudflare, Inc.	Frontend hosting and CDN (Cloudflare Pages)	IP addresses, request metadata, static page delivery	Global (Cloudflare edge network)
Railway Corp.	API backend hosting	API request payloads, response data, application logs	United States
Clerk, Inc.	User authentication and account management	Email address, hashed passwords, session tokens	United States
Stripe, Inc.	Payment processing and subscription management	Billing information, card data (PCI-DSS compliant vault), transaction history	United States
Resend, Inc.	Transactional email (key issuance, receipts, notices)	Email address, name, email content	United States
Anthropic, PBC	AI-assisted evidence synthesis (evidence grading pipeline)	Ingredient name, outcome, anonymised corpus excerpts. No personal data is sent to Anthropic.	United States

Subprocessors

Evidence Data Source

Subprocessor Change Notification