All projects matching “llm”

Is this EU VAT number valid? Live VIES check against the European Commission's official VAT Information Exchange System — validate a European VAT id, verify a business tax id before invoicing, confirm B2B reverse-charge / zero-rating eligibility. Returns the registered trader name and address where the member state discloses them. Deterministic: no LLM. Distinguishes a genuinely invalid number fro

Use when a page will not read correctly with a plain fetch — a single-page app, content loaded by JavaScript, or markup so noisy the text is buried. Fetches it with a headless browser (rendering JavaScript) and returns the main readable content as clean Markdown — ideal to feed an LLM as context. JSON body { url } (public http/https), optional include_images (default false). Returns title, final_u

Answers questions (single-turn question answering) from a small local LLM (qwen3.5 2B). Best for general knowledge, definitions and simple reasoning. POST JSON {"input": string}, up to 3000 characters; returns {answer, model, elapsed_ms}. Answers are brief (2-3 sentences). Typically 5-15 s. Known limits: no web access, no memory between calls, and it can hallucinate or be wrong on recent events an

Metalift scrape: URL → LLM-ready markdown with WAF/retail escalation, JS render, and residential proxy when needed. Use instead of a plain HTTP fetch when Cloudflare, bot checks, or SPAs block cheap scrapers. Also works for static pages (pay only for the path used). (5 credits ≈ $0.005 USDC)

Rail/coach/ferry/cruise compensation eligibility — UK Delay Repay, EU Rail 2021/782 (25/50%), EU/UK ferry 1177/2010, EU coach 181/2011, US FMC cruise refunds (46 CFR 540, 3-day rule), Athens Convention ship baggage — plus honest answers for VIA, Amtrak, India TDR, Japan JR. Deterministic percentages, article-level citations, claim-deadline math, operator defenses. No LLM.

ClinicalTrials.gov lookup, deterministic. Answers "what is the status of NCT02142738", "recruiting phase-3 melanoma trials", "Pfizer trial pipeline". Single trial by NCT id (status, phase, sponsor, enrollment, dates, countries) or search by condition/sponsor with phase and status filters. Typed records straight off the official v2 API — no LLM, no key, public domain.

[MCP: optimize_context] Pay-per-call JSON context pruning & LLM token budget enforcement. Saves up to 85% prompt tokens.

SWARM Read Page - Retrieve the content of a URL as clean, LLM-ready text (headings with #, lists with -; menus, scripts, styles and markup removed). Use when an agent needs to read a web page, article or docs page without paying model tokens for raw HTML: on 15 public pages we measured, the text was a median 97% smaller than the HTML (about 67,000 fewer tokens per page). Requires `url`; HTML or te

Bibliography verification for AI research agents — POST up to 20 citations (DOIs or free-text references) and get a per-reference verdict: exists/not-found, retracted/withdrawn (Crossref + Retraction Watch + OpenAlex), and title/author/year mismatches against a real record. Built for agents auditing LLM-generated bibliographies before a citation ships.

Live grid carbon intensity, deterministic. Answers "how clean is the grid right now", "gCO2 per kWh in Germany", "cleanest hour to run a job today". Current intensity + forward curve (to 48h) + cleanest and dirtiest slots, from NESO (GB national + 14 DNO regions) and energy-charts/Fraunhofer ISE (DE FR NL BE AT ES IT NO). No LLM, no key. Pair with /api/energy/compute-window for full job scheduling

Eight quarters of a US public company's reported fundamentals, straight from SEC XBRL. Answers "Apple revenue last 8 quarters", "is NVDA net income growing", "MSFT EPS trend". Revenue, net income, diluted EPS per quarter with computed YoY changes and the exact XBRL tag used — every number verbatim from data.sec.gov companyfacts, nothing modeled. Accepts ticker, company name, or CIK. No LLM, no key

Live USGS river-gauge tick, deterministic. Answers "what is the flow at this gauge right now", "is this river high or low vs this week". Returns the latest streamflow (discharge, cfs) and gage height (ft) for one USGS site plus the 7-day min/max range, reading count, site name and coordinates, straight from USGS instantaneous values. Unknown sites 404 before payment. No LLM, no key, public domain.

CVE fact check, deterministic. Answers "is CVE-XXXX-NNNN real", "is it exploited in the wild", "what version fixes it", "how severe is it". Joins the NVD record (CVSS score, vector, dates), the CISA KEV catalog (actively-exploited flag, remediation due date, ransomware use), and OSV (affected packages + fixed versions per ecosystem). No LLM. Distinguishes "not a known CVE" from "registry unreachab

Extract clean, LLM-ready Markdown or plain text from up to 5 public webpage URLs. Removes page boilerplate and returns structured per-URL results for AI agents, RAG ingestion, research and summarization.

Federal Register search, deterministic. Answers "new SEC rules on stablecoins", "is there an open comment period on AI", "when does this rule take effect". Full-text search over every US federal rule, proposed rule and notice, with effective dates, comment deadlines and docket ids — or mode=open-comments for comment periods closing soonest. Typed documents straight off the official API. No LLM, no

Reader Pro: fetch a public web page and return structured data, no LLM: OpenGraph/meta tags, parsed JSON-LD (schema.org), HTML tables as row objects, deduped internal/external links, emails/phones/social profiles, heading outline, images with alt text, plus optional CSS-selector fields.

LLM chat completion with no API key or signup (fast open model, meta-llama/llama-3.3-70b-instruct). OpenAI-compatible: set your SDK base URL to https://keyless.run/fast/v1. Flat $0.002 per call, up to 1500 output tokens. Charged only on success.

US safety recalls for a VIN, deterministic. Answers "does this car have open recalls", "is this used car safe to buy", "what NHTSA campaigns cover this vehicle". Decodes the VIN via NHTSA vPIC (year, make, model, body, engine) then lists every NHTSA recall campaign for that vehicle: component, summary, remedy, park-outside and do-not-drive flags. No LLM, no key, public domain.

Fetch any public URL and return its content as base64 with the content type and a ready-to-use data URI. Embed images, PDFs and files directly in LLM and API calls without downloading them yourself. Clean JSON.

Flight compensation eligibility across 7 regimes — EU261, UK261, Canada APPR (incl. s.20 denied boarding), Brazil ANAC 400, Turkey SHY, India DGCA, Israel ASL — plus Montreal Convention baggage claims (effective-dated SDR caps, Art. 31 deadline math): deterministic amounts, legal citations, carrier defenses, evidence list. No LLM — every number is recomputable. Eligible? Get the ready-to-send clai

Stablecoin peg tick — USDT, USDC, USDS, DAI, USDe, USD1 in ONE $0.01 call: live price, deviation from $1, 24h change, and a deterministic STABLE / WATCH / ALERT verdict per coin. Built for monitoring loops: call it every cycle, escalate to /api/depeg-watch for the single-coin deep read when a verdict is not STABLE. 60s freshness, cited to CoinGecko, no LLM in the path.

Returns a 0–100 generative-engine visibility score and estimated LLM mention rate for a URL. Cost $0.01 per call. Not HTTP uptime, TLS, or a credit.

Fetch any public web page or PDF and return clean Markdown, plus title, description, author, publish date, JSON-LD and outgoing links. Built for LLM agents.

Ultra-processed food check for one real product — scan a barcode (EAN/UPC) and get the NOVA processing group, Nutri-Score, full additive census with flagged-additive notes (EU-banned dyes, IARC-listed sweeteners, nitrites), added sugar/salt/saturated fat per 100g, and a plain verdict. Live Open Food Facts data (3M+ products, global), deterministic, no LLM.

Base token risk check before a swap: send a token contract address, get ok / caution / avoid with reasons. Reads the contract (ERC-20 metadata, owner, upgradeable proxy, mint / pause / blacklist / fee selectors in bytecode) and its DEX liquidity (Uniswap + Aerodrome pools, read on-chain). Deterministic, no LLM.

URL-to-Markdown web extraction and context compression for AI agents. Fetch a public web page and return clean, task-relevant, citation-verifiable LLM context with exact excerpts, locators, hashes, and estimated tokens saved. Use before an agent summarizes, answers questions from, or cites a long or noisy URL. Deterministic and model-free. Public HTML only; no PDFs, logins, or JavaScript-only page

Fetch a public URL and return clean LLM-ready markdown from static HTML.

AI Chat (LLM) chat: OpenAI-compatible chat completion on Llama 3.3 70B or Qwen 2.5 Coder 32B (up to 32000 characters in, 1024 tokens out, optional JSON mode).

Use when an agent has the text of a Japanese invoice or receipt (e.g. OCR output) and needs exact amounts without LLM misreads. Returns registration number (T+13, check digit only, not an NTA registry lookup), total, subtotal, 8%/10% tax breakdown and issue/due dates (Reiwa to ISO). Deterministic rules; text only, no images/PDF. Errors before payment (no charge): 400 body not {text} or over 5000 c

Verify up to 50 citations in one call, one flat price. Body: {"citations": ["<string>", ...]} - same shapes and verdicts as the single route. Returns results[] (one per citation), the verified/not_found/suspicious counts, verification_time_ms, hallucination_risk (LOW all verified, MEDIUM 1-2 not found, HIGH 3+). Deterministic, no LLM in the path. suspicious = a name_match mismatch: a DIFFERENT cas