Financial Document Intelligence API
filedge parses 10-Q, 10-K, 8-K, Form 4, and earnings call transcripts into clean JSON. Pay $0.05 per request in USDC. No API keys. No accounts. Built for agents.
Used by autonomous agents · x402 v2 · Base Network · FinBERT-powered NLP
POST https://filedge.io/v1/parse
{
"source": "edgar",
"type": "10-Q",
"ticker": "AAPL",
"period": "latest"
}
→ Returns structured JSON in ~200ms
How It Works
x402 handles payment inline. Your agent sends a request, pays $0.05 USDC, and gets structured data back. That's the entire flow.
POST to /v1/parse with the ticker, document type, and period. No API key. No authentication header.
POST /v1/parse
{ "source": "edgar", "type": "10-Q", "ticker": "NVDA" }
The API returns HTTP 402 with payment instructions. Your agent pays $0.05 USDC on Base using the x402 protocol. No wallet setup required in your code — the x402 client library handles it.
The API returns typed JSON. Financial values in USD cents (integer). Flags, signals, and metadata included. Cache-aware — repeated requests don't re-parse.
API Reference
POST /v1/parse accepts SEC EDGAR filings by ticker and earnings call transcripts. Returns typed JSON every time.
edgar | transcript | url
10-Q | 10-K | 8-K | form4
e.g. AAPL
latest (default) | YYYY-QN | YYYY
Direct URL to filing
{ "meta": { "ticker": "AAPL", "company": "Apple Inc.", "document_type": "10-Q", "period_end": "2024-12-28", "filed_date": "2025-02-07", "parsed_at": "2025-04-22T18:04:11Z", "cache_hit": false, "source_url": "https://www.sec.gov/Archives/edgar/data/320193/..." }, "financials": { "revenue": { "value": 12430700000, "currency": "USD", "yoy_pct": 4.1 }, "net_income": { "value": 3620600000, "currency": "USD", "yoy_pct": 7.9 }, "eps_diluted": { "value": 2.40, "yoy_pct": 10.1 }, "gross_margin_pct": 46.9, "operating_income": { "value": 3820000000, "currency": "USD" } }, "segments": [ { "name": "iPhone", "revenue": 6924500000, "yoy_pct": 1.1, "pct_of_total": 55.7 }, { "name": "Services", "revenue": 2624900000, "yoy_pct": 13.9, "pct_of_total": 21.1 }, { "name": "Mac", "revenue": 897000000, "yoy_pct": 15.6, "pct_of_total": 7.2 } ], "guidance": { "provided": true, "revenue_range": { "low": 88500000000, "high": 91500000000 }, "signal": "raised" }, "risk_flags": [ { "text": "Significant exposure to China revenue concentration risk", "severity": "high" }, { "text": "Foreign exchange headwinds may impact international margins", "severity": "medium" } ] }
Quarterly Report
Returns
Revenue, net income, EPS, segments, guidance, risk flags
Annual Report
Returns
Full-year financials, MD&A highlights, risk factors
Material Event
Returns
Event type, key disclosures, officer changes, earnings releases
Insider Transactions
Returns
Transactions list, net USD bought/sold, signal (buying/selling/mixed)
Earnings Call
Returns
Tone scores, key topics, forward signals, analyst Q&A quality
Pricing
No monthly fees. No tiers. No rate limit tiers. No account required. Each request costs exactly five cents, paid in USDC on Base. Your agent pays as it goes.
Price per request
Agents pay automatically via the x402 client library. No billing portal. No invoices. The chain is the receipt.
View x402 docs ↗Built on Solid Ground
No magic. No hallucinations. Structured extraction from primary sources, with NLP signals layered on top.
All 10-Q, 10-K, 8-K, and Form 4 data comes directly from the SEC's EDGAR XBRL feed. No third-party data vendors. No re-licensing risk. Primary source, every time.
Transcript tone and risk flag sentiment scoring uses FinBERT — a financial-domain BERT model trained on financial news and filings. Not a general-purpose model repurposed for finance.
Payment uses the open x402 standard, implemented by Coinbase. Your agent's x402 client library works with any x402 service, not just filedge. No vendor lock-in on the payment layer.
Parsed documents are cached in Redis. Requesting the same filing twice doesn't cost twice. Cache TTLs are set by document type — shorter for 8-K (volatile), longer for 10-K (stable).
Monetary values are integers (USD cents) — no floating point currency. Dates are ISO 8601. Enums are strings with fixed vocabularies. Every field is typed and predictable.
Exposes /.well-known/x402.json for service discovery and /llms.txt for AI agent instructions. Your agent can find, understand, and use this API with zero human setup.
Python · Node.js · Any x402-compatible agent · FinBERT NLP · Redis cache · Base Network · EDGAR XBRL