Straight answers.

Three sources, two maintained by the SEC and one for transcripts:

• SEC EDGAR: The full text of every 10-K, 10-Q, and 8-K.
• SEC XBRL company-facts API: Structured financial figures submitted directly by the filers.
• Earnings call transcripts: Sourced from SEC filings where available, company IR pages, live transcription, and third-party vendors. Quality is generally reliable for US large and mid caps.

Three filing types are fully supported today:

• 10-K: Annual reports. Full risk-factor analysis and MD&A.
• 10-Q: Quarterly reports. Same structure, shorter horizon.
• Earnings calls: Live Q&A transcripts. Q&A evasion detection on management's answers, where executives deflected, hedged, or refused to address what analysts pressed on.

We also monitor 8-Ks. When one is published for a ticker on your watchlist, Sharpread analyzes it automatically and emails you the event type, severity, and a verbatim quote.

Sharpread polls EDGAR once each weekday at 4pm ET. We email you within minutes of detecting a new filing for any ticker on your watchlist. A filing made earlier that day usually reaches your inbox by 4:30pm ET. Overnight or weekend filings are caught in the next weekday's poll.

8-K alerts run through the AI engine first. They arrive enriched with severity, a signal (Bullish, Bearish, Neutral), and a verbatim quote. 10-K and 10-Q alerts link straight to the filing so you can run the full analysis when you are ready.

Four main differences:

• Management language analysis. Sharpread extracts and classifies hedge language, defensive framing, and conspicuous omissions. ChatGPT just generates generic summaries.
• XBRL numbers. Hard financials come from SEC structured data. ChatGPT cannot read XBRL and will guess numbers from prose. We built this platform specifically to prevent that hallucination.
• Validated prompts. Our prompts are stress-tested across healthy, struggling, and distressed companies (AAPL, INTC, BYND). We calibrated the outputs to eliminate false-positive red flags on healthy companies.
• Structured output. Every brief has the exact same fields: bull case, bear case, red flags, and a verdict. This lets you actually compare filings side-by-side.

Sharpread automatically compares the new brief against your prior one for the same filing type. You see exactly what shifted in management tone, which red flags resolved, and how guidance changed.

The QoQ Intelligence Dashboard surfaces these trends across every ticker you have analyzed. You get a signal timeline, tone history, and the underlying XBRL metrics side-by-side.

Yes. When analyzing a ticker, you can use the dropdown to backfill up to 12 recent filings in a single batch. This builds your quarter-on-quarter trend data instantly. Each filing in the batch consumes one analysis credit, regardless of whether it hits our shared cache.

Sharpread covers any company that files with the SEC, including US-listed ADRs of foreign issuers (TSM, BABA, NVS, NVO). That covers roughly 8,000 tickers across the NYSE, Nasdaq, and OTC markets.

Companies that file only in their home jurisdiction are not supported. Our data pipeline is SEC EDGAR-only. Foreign private issuers that file only 20-F or 6-K are also not covered yet.

If a ticker fails, check the spelling first. Brand-new IPOs sometimes take a few days to appear in EDGAR. Email hello@sharpread.ai if you are unsure about a specific ticker.

One analysis equals one filing for one ticker. Running AAPL's 10-K counts as 1. Running AAPL's 10-Q is 1 more. Loading a past analysis from your history does not count.

Cache-accelerated analyses still count. If another user recently analyzed the same filing, Sharpread serves you the result instantly from our shared cache. That still counts against your quota since you consumed the brief. The cache saves us the cost of a fresh Claude call, but the value to you is the same.

No. Top-up credits carry over indefinitely. If you buy a $5 top-up (15 credits) and only use 10 this month, you start next month with 5 unused credits plus your fresh 50 monthly allowance.

For Pro subscribers who regularly run historical backfills, hitting your monthly limit can interrupt your workflow. You can opt in to Auto-Topup from your dashboard settings. If enabled, we will automatically charge $5 for 15 more analyses when you hit your limit. It is strictly opt-in (default OFF), capped at a maximum of 3 auto-topups per month to prevent runaway billing, and you can disable it with one click at any time.

Yes. Stripe manages all subscriptions. You can cancel from your account at any time and keep Pro access until the end of your current billing cycle. Top-up credits you already purchased do not disappear. They stay on your account even if you cancel and resubscribe later.

No. Sharpread generates AI analysis of public SEC filings. It is a research tool, not a regulated advisor or personal investment advice.

The hard financial numbers in every brief (revenue, net income, EPS, total assets, total debt) come directly from the SEC's XBRL structured-data API. These are the exact numbers the company submitted. They are not generated by AI and cannot be hallucinated.

The qualitative analysis (management tone, red flags, verdict) is produced by an AI model. Every brief carries a strict disclaimer to verify against the source documents before acting. Use it to inform your research, not replace it.

Every analysis is private by default. They are saved to your account and visible only to you. The public analysis URL returns a 404 to anyone else until you explicitly choose to publish it.

Clicking SHARE on the analysis page publishes it and copies the URL. Only then can you send the link to someone and have them see the brief. You can publish individual analyses at any time. There is no setting that flips your whole account public.

The Trending feed shows which tickers are being analyzed but never links back to individual accounts.