Frequently Asked Questions

Pricing & quotas

Q1. What does a Report cost?

Each Report counts as one unit against your plan's monthly quota — there's no per-Report dollar charge and no premium tier that costs more. Your plan grants a fixed number of Reports each month; see the Pricing page for current limits. Quotas reset on each billing cycle.

Q2. Do unused units roll over?

No. Monthly quotas reset on each billing cycle. Refines reset daily at 00:00 UTC.

Q3. Is there a premium or "Pro" tier?

No. Every Report — on every plan — is generated with the same multi-model synthesis and adjudication. Plans differ by how many Reports you get each month (and team features), not by output quality. You never have to choose a tier or pay more for a "better" engine.

Q4. Is there a limit on how much I can input?

Yes. Your prompt plus any attachments and Vault context must stay under the input limit (about 250,000 tokens); oversized requests are rejected before they run. If a request is large but within range, the system may trim lower-priority source material mid-flight to stay within cost limits, and shows a banner on the Report when it does.

Q5. What if my Report fails?

Your quota is automatically refunded in full. The Reports page will show the failure with a brief error reason. Common causes: backend timeout, model API outage, public source unavailable.

Q6. What if I cancel mid-month?

Cancellation takes effect at the end of your current billing period. You keep access and your remaining quota until then. We don't pro-rata refund partial periods.

Q7. How do I change plans?

Settings → Billing → Open billing portal → "Update subscription". Upgrades take effect immediately with pro-rata billing. Downgrades take effect at next renewal.

Q8. Do you offer annual billing?

Not yet. We're prioritising product features first. If you need annual billing for procurement reasons, contact sales@essence.report.

Q9. Is there a free trial?

We don't offer free trial Reports. Before subscribing you can:

Use Refine (200/day) — see how we tighten your questions
Browse sample reports — illustrative outputs that show the format

A paid plan is required to generate your own Reports.

Reports & quality

Q10. How long does a Report take?

Typically 5–10 minutes. Some queries with heavy source retrieval can take longer. Hard limit is 30 minutes — if a job hasn't completed by then, the system auto-fails and refunds your quota.

Q11. Can I see the Report being generated live?

The progress page shows an estimate. For privacy and IP reasons, we don't expose live model output during generation. The full Report appears when complete.

Q12. Why are some Reports marked "Complete with warning"?

Common reasons:

Truncation: the adjudicator hit its output limit; the Report is mostly complete but may end abruptly
Auto-scaled: an oversized request was automatically scaled mid-flight to stay within cost limits
Source retrieval partial: web crawling was truncated to fit input limits

Each warning is shown as a banner on the Report page.

Q13. Are the citations real?

Yes. Citations to web URLs are gathered by our retrieval pipeline at the time of your query. We do best-effort verification that URLs are valid at gathering time. URLs may go stale; we don't continuously re-verify after delivery.

Q14. Can I trust the conclusions?

Reports are research artifacts, not professional advice. They're designed to give you a multi-model perspective with verified open-source evidence. Always verify critical findings independently before relying on a Report for decisions.

Q15. The Report says something I know is wrong. What do I do?

AI models can hallucinate, especially on niche topics, very recent events (last 24 hours), or topics with sparse online coverage. Best practices:

Cross-check key facts against the cited sources
For factual disputes, refine your prompt to ask the system to "cite primary sources only" and re-run

Q16. Can I get the Report in a specific format?

The default is Markdown. From the Report page, you can export to:

Markdown (.md) — most portable
PDF (.pdf) — with optional account watermark
Word (.doc) — opens in Word, Google Docs, LibreOffice

You can also ask in your prompt: "respond as a table" / "respond as bullet list" / "respond in legal memo format".

Privacy & security

Q17. Do you train AI models on my data?

No. We do not use your prompts, uploaded files, Vault entries, or generated Reports to train our own or third-party AI models. See § 4.4 of the Terms and § 3 of the Privacy Policy.

Q18. Where is my data stored?

Backend processing: Vultr Tokyo region
Generated Reports: 7-day TTL in our Tokyo backend
Account data: Clerk (US-based, GDPR-compliant)
Billing: Stripe (PCI-DSS Level 1)
Caching: Upstash (configurable region)

See Privacy Policy § 5 for full processor list.

Q19. How long are Reports kept?

7 days on our backend, then auto-deleted. To keep:

Export immediately (Markdown is most portable)
Pin to your local browser-side list

Materials you save to your Vault persist for reuse across Reports (separate from the generated Reports above, which still expire after 7 days).

Q20. Is my prompt visible to the LLM provider?

Yes — generating a Report sends your prompt to several third-party large language model providers, each with its own data-handling policy. We disable training-data retention where a provider offers that option, but cannot guarantee zero retention at every upstream provider. We never use your data to train our own models.

Q21. Can I use Essence.Report for confidential / regulated data?

For most professional research use cases: yes. For data subject to strict regulatory regimes (HIPAA PHI, EU Article 9 special-category data, classified information), contact sales@essence.report to discuss enterprise arrangements including DPA, data residency, and processor restrictions.

Q22. Is two-factor authentication available?

Yes. Settings → Security. We strongly recommend enabling it.

Service operations

Q23. What if the service is down?

We're a startup — we don't yet offer 99.9% SLA. Our uptime monitoring is at status.essence.report (link forthcoming). If you experience an outage, your in-flight job will either complete or auto-refund quota.

Q24. Can I generate Reports programmatically via API?

Not in MVP. We're focused on the web product first. Enterprise customers can discuss API access with sales@essence.report.

Q25. Do you support file types other than PDF, Word, Markdown, TXT?

Not yet. CSV, XLSX, JSON, and image-based PDFs (with OCR) are on the roadmap. For now, convert to one of the supported formats before uploading.

Q26. Is there a mobile app?

The web product is fully responsive — works on phone and tablet. A native app is not planned for MVP.

Q27. What languages does the system speak?

The LLMs handle English plus most major languages. Sample Reports and our marketing are currently English-only. The interface is English-only at MVP; we plan zh-TW and other languages later.

Account & teams

Q28. How do I invite teammates?

For now, Team admin actions (adding seats, adjusting per-seat caps) go through support@essence.report — email us your Team ID and the changes. Self-service team admin is on the roadmap.

Q29. Can two people use the same account?

No. Account sharing is a violation of our Terms (§ 4). Each user needs their own account; subscribe to a Team plan if you need multi-user access.

Q30. How do I delete my account?

Email privacy@essence.report with the subject "Account deletion request" from the email address associated with your account. We'll process within 30 days, subject to legal retention requirements (e.g. tax records).