ExtractMarkdown
Scrape any URL and convert the page to clean Markdown. Strips nav, footer, scripts, and non-essential HTML to minimize tokens. SSRF-protected with a 5MB page limit.
VerifyEmail
Check if an email address is disposable or from a known burner domain. Validates format, checks against a curated domain list, and verifies MX records via DNS.
ResolveTimezone
Convert latitude/longitude to full timezone data including IANA timezone, UTC offset, DST status, current local time, and next transition.
InferSchema
Auto-detect the format (JSON, CSV, XML) of a payload and infer its schema — field names, types, nullability, and nesting.
GenerateProductSchema
Generate JSON-LD Product structured data for SEO. Outputs both raw JSON-LD and a ready-to-embed <script> tag.
CleanCsv
Clean, normalize, and deduplicate CSV data. Fixes encoding, normalizes headers, collapses whitespace, pads/truncates rows, and optionally normalizes date formats.
CloakLink
Create cloaked UTM-tagged tracking URLs. Encode campaign parameters into a short token and decode them back.
RepairJson
Fix broken JSON from LLM outputs — trailing commas, unquoted keys, single quotes, missing brackets. Optionally validate against a JSON schema.
CountTokens
Count tokens for multiple LLM models (GPT-4, Claude, Llama, Mistral, Gemini). Estimates costs, context window usage, and flags when text exceeds the model's limit.
GradeEmailAuth
Audit a domain's email authentication setup: SPF, DKIM, DMARC, MTA-STS, and BIMI. Returns a letter grade (A–F), score, and actionable recommendations.
RedactPii
Detect and redact personally identifiable information from text. Handles emails, SSNs, credit cards, phone numbers, IP addresses, street addresses, dates of birth, and URLs.
ValidatePhone
Validate international phone numbers across 30+ countries. Returns E.164 format, country info, number type (mobile/landline), and formatted output.
ValidateTaxId
Validate VAT and Tax IDs for EU, US, UK, Australia, and India. EU VAT numbers are verified against the VIES service when available.
DetectPromptInjection
Score text for prompt injection and jailbreak attempts. Uses pattern matching and heuristic analysis across multiple attack categories. Returns a risk score (0–100) with flagged spans.
SemanticSearch
Match a search query against a catalog of items using TF-IDF similarity. Supports fuzzy matching, synonym expansion, and configurable title weighting.
FindRelatedContent
Find related content items from a corpus using TF-IDF and Jaccard similarity. Configurable thresholds, category boosting, and optional score breakdowns.
ParseTasks
Extract actionable tasks from voice transcripts. Identifies action verbs, assignees, priorities, deadlines, and confidence scores. Handles up to 100,000 characters.