vaaya.ai

Firecrawl — Extract structured data from URLs using a schema.

Firecrawl — Crawl a website starting from a URL, following links.

Modal — Create a sandboxed compute environment (CPU by default, 300s timeout). Pass `gpu` (T4 | L4 | A10G | A100 | H100) for GPU workloads like model inference.

Firecrawl — Scrape a single URL and return clean markdown/HTML.

Stage a media file on the fal CDN before a fal generation. Returns { upload_url, file_url }: PUT the raw file bytes to upload_url (with the matching Content-Type header), then pass file_url as the image_url/video_url/audio_url param of the generation. Upload originals UNCOMPRESSED — never downscale or re-encode; high-res inputs cost nothing extra and compression degrades outputs.

Parallel — AI-powered web search via x402 (1¢ flat).

Exa — retrieve content for URLs or document IDs via our API key. Charges 0.1¢ per (url or id) × content field requested.

Apify — Tweets by search term, handle, or conversation.

TikHub — 742 per-call social-data endpoints (GET) across douyin/tiktok/weibo/instagram/bilibili/zhihu/kuaishou/youtube & more. Pass { endpoint, ...params } (params per endpoint; most reads take username/user_id/keyword + pagination_token). Find endpoints via vaaya/discover or the social-data knowledge doc. Most calls 1¢.

Exa — semantic web search via our API key. numResults up to 100. For people/lead discovery set category:"people" with an open ICP query.

Modal — Run a command in a running sandbox and return its output.

Parallel — Start an async AI research task. Pricing depends on `processor`: pro 10¢, ultra 30¢. Returns { run_id }; poll task-status until completed.

Firecrawl — Search the web and return scraped results.

Generate or edit images, video, music, and speech via fal.ai. Pass `model` (one of the registered keys) plus the model-specific params (prompt, image_url, aspect_ratio, duration, text, etc.).

Reveal a person's verified email via Prospeo. Pass `person_id` (from search-people) or `linkedin_url`. Returns the unmasked verified email. 1 credit; free if the record was enriched within 90 days.

Firecrawl — Map all URLs on a website without scraping content.

Paid x402 endpoint for newly created prediction-market data and emerging betting topics.

Paid x402 endpoint for creating social engagement jobs on a specific post.

Apify — Recent posts from LinkedIn profile or company URLs.

ScrapingAnt — Scrape a URL and return LLM-ready markdown (rendered in headless Chrome, then converted). Pass `url`; optional browser/wait_for_selector/proxy_country/timeout knobs. Returns `{ url, markdown }`. Alternative to crw/scrape or firecrawl/scrape when you want ScrapingAnt's proxy pool behind the fetch.

Fetch one SEC filing document from EDGAR Archives by { cik, accession, filename } (from edgar/fulltext hit ids or edgar/filings). Returns the raw document — Form D XML carries offering amounts, Form C/C-AR XML carries revenue/net income of private companies, Form 4 XML carries insider buys/sells. Prefer .xml/.htm/.txt files (not PDFs).

Linkup — DEEP agentic search: iterative multi-query retrieval for hard or multi-hop questions where one-pass search misses. Same params as linkup/search (`q` + outputType/dates/domain filters). Slower (tens of seconds) and 5× the price — escalate here only after linkup/search or exa/search come back thin. For full multi-step research reports, parallel/task (10-30¢) is the bigger rung.

ScrapingAnt — AI data extraction WITHOUT a schema: describe the fields in plain English and get structured JSON back. Pass `url` and `extract_properties` — a comma-separated list like 'product title, price(number), full description, reviews(list: review title, review content)'. Optional browser/wait_for_selector/proxy_country/timeout knobs. Vendor cost scales with page length (1 credit per 30 chars of markdown) — prefer crw/extract (schema-driven, 5¢) for big pages; this rung is for quick no-sc…

Tavily — AI-native web search tuned for RAG: ranked results with relevance scores, optional LLM answer synthesis. Pass `query`; optional `topic` (general|news|finance), `max_results` (1-20), `include_answer` (true = synthesized answer), `include_raw_content` (true|"markdown" = full page text per hit), `include_domains`/`exclude_domains`, `time_range` (day|week|month|year), `country`. Alternative to exa/search (semantic recall) and brave/search (keyword index).

Find a person's professional email via Tomba from their name + company. Pass `domain` (preferred, e.g. stripe.com) or `company`, plus `full_name` or `first_name` + `last_name`. Returns the most likely email with a confidence score, sources, and position data. 1 Tomba credit — only charged upstream when a valid email is found.

Search federal court dockets via RECAP (CourtListener v4) — the litigation profile of any company: every case naming it, parties, attorneys, filing dates, live docket links. Same params as courtlistener/cases. Coverage is federal courts via the RECAP archive; state-court coverage varies — say so in outputs.

Search all US nonprofits by name/keyword (ProPublica Nonprofit Explorer, IRS 990 data). Pass `q` + optional `state` (2-letter) and `page`. Returns orgs with EINs — feed an EIN to propublica/nonprofit for financials. The route into any 990: foundations, research labs (early OpenAI), the SF NGO network.

EDGAR daily index: EVERY filing of EVERY form type for one day (plain-text form.idx). Pass `date` (YYYY-MM-DD, business days; weekends/holidays 404 = no filings). Rows: form type, company, CIK, date, filing path. THE way to enumerate all Form Ds (or any form) in a date window — one call per day; filter rows by form prefix; full-text/kadec0 search cannot do this (they need a query term).

SEC EDGAR full-text search over all filings (2001+). Pass `q` (quote for exact phrase) + optional `forms` (comma-separated: D = private fundraises, C/C-AR = Reg CF financials, 1-K/1-SA = Reg A financials, 4 = insider trades, UPLOAD/CORRESP = SEC comment letters), `startdt`/`enddt` (YYYY-MM-DD), `from` (pagination offset). Hits include the filer, date, and {accession}:{filename} ids for edgar/document.

Kadec0 — 29 public-data endpoints over x402: academic papers, CVE, FDA/recalls, SEC EDGAR, congress trades, USGS quakes, permits, whois/domain/DNS, geocode, weather, nutrition, HackerNews/StackOverflow, DeFi TVL/yields/gas. GET { endpoint, ...query }. 1-21¢ caps. See research + onchain-data knowledge docs.

TheirStack — Search job postings across thousands of career sites and job boards (hiring signals, recruiting, market research). REQUIRED: `limit` (1-10, vendor bills per job returned) plus at least one of `posted_at_max_age_days`, `posted_at_gte`/`lte`, `company_domain_or`, `company_linkedin_url_or`, `company_name_or`. Rich filters: `job_title_or`, `job_description_pattern_or` (regex), `job_technology_slug_or` (case-sensitive slugs — resolve via theirstack/tech-catalog), `job_country_code_or` (…

TheirStack — List the technologies a company uses, each with confidence (low/medium/high), the number of job posts mentioning it, and first/last dates seen. Pass ONE of `company_domain` (preferred), `company_name`, `company_linkedin_url`; optional `keyword_category_slug_or`, `confidence_or`, `min_jobs`, `limit` (rows). Flat vendor cost per company lookup regardless of rows; free upstream when the company isn't found.

Modal — Terminate a running sandbox.

Apify — Indeed job listings by role title.

CRW — Scrape a single URL to clean markdown/HTML/JSON (Firecrawl-compatible). Pass `url`; optional `formats` (markdown|html|rawHtml|plainText|links|json|summary|screenshot, default markdown), `onlyMainContent` (default true), `renderJs` (null = auto-detect), `waitFor` (ms), `stealth` (anti-bot browser), `country` (2-letter proxy egress), `jsonSchema` (structured output). Sync — returns the scraped content directly. Cheapest full-featured scrape rung.

CRW — Discover the URLs of a website without scraping content (sitemap + crawl fallback). Pass `url`; optional `maxDepth`, `useSitemap` (default true), `timeout` (seconds, default 120). Recon step before targeted crw/scrape or crw/crawl.

Brave — news-only search over the Brave index: recent articles with source, age, and breaking flags. Pass `q`; optional `count` (1-50), `country`, `search_lang`, `freshness` (pd|pw|pm|py), `safesearch`, `offset`. Use for current-events queries where web search returns stale evergreen pages.

CRW — Search the web and optionally scrape the hits in one call. Pass `query`; optional `limit` (1-20, default 5), `tbs` (freshness: qdr:h|d|w|m|y), `sources` (web|news|images), `lang`, and `scrapeOptions` ({ formats, onlyMainContent }) to return full page content per hit instead of snippets.

Verify an email address's deliverability via Tomba. Pass `email`. Returns status (deliverable/undeliverable/risky), SMTP + MX checks, disposable/webmail flags, and a score. Cheapest verify rung — use before sending to a found email.

One nonprofit's full IRS 990 history by EIN (ProPublica): year-by-year revenue, expenses, officer compensation, assets, plus links to the filed 990 PDFs. The primary source for "what does this NGO/foundation actually take in and pay out".

SEC EDGAR filing history for one company by CIK (data.sec.gov submissions). Returns entity metadata + every filing (form type, date, accession number) — the sweep source for a company's Form Ds (fundraises), Form 4s (insider trades), 8-Ks, S-1s. Find the CIK via edgar/fulltext hits or the company's EDGAR page.

Tavily — extract clean page content (JS handled) from up to 5 URLs you already have, in one 1¢ call. Pass `urls` (array, ≤5); optional `format` (markdown default | text). For bigger batches make multiple calls. Alternative to exa/contents (0.1¢/url, no JS) and firecrawl/scrape (1¢/url, full rendering).

SEC EDGAR entity search: company/fund name → registrant CIKs (the autocomplete index). Pass `q` (name fragment). Returns ranked entities with CIK. START HERE for any company-scoped EDGAR question — resolves brand → legal entity, finds SPV families, and proves "never registered" negatives that full-text search cannot.

Search 250M+ scholarly works (OpenAlex — successor to Microsoft Academic Graph). Pass `search` and/or `filter` (comma-joined, e.g. from_publication_date:2024-01-01,is_oa:true,cited_by_count:>100), optional `sort`, `page`, `per_page`. Returns metadata + abstracts + citation counts + OA links. The citation-graph layer kadec0/papers lacks.

TheirStack — Search the catalog of tracked keywords: technologies AND buying-intent topics. The slug resolver for every other theirstack filter (slugs are case-sensitive). Pass `q` (free-text search); optional `keyword_type` ("technology" = techs only), `keyword_type_not` ("technology" = intent topics only), `category_slug`, `limit` (default 25). Free upstream — always resolve slugs here before filtering jobs/companies by technology.

Search 10M+ US court opinions (CourtListener v4). Pass `q` + optional `court` (e.g. cand, ca9, scotus), `judge`, `filed_after`/`filed_before`, `order_by` ("dateFiled desc"), `cursor` (pagination). Returns case names, courts, dates, cites, and absolute_url links to full opinions.

Entity-resolved, de-duped company/industry news with sentiment, AI summary, and tagged company mentions (~80% of noise filtered). Scope by { company: 'https://canva.com' } (website URL), { industry } (taxonomy code from akta:industry-search), or free-text { query }. Optional filters (additive): `sentiment_list` (positive/negative/neutral), `type_list` (tag CODES, e.g. CM01=Equity Fund-Raising, WT01=Strategic Hiring, SD01=Product Launches, SD02=Partnerships — codes appear on returned articles' `…

Browserbase — Create a headless browser session via x402. Charges 0.2¢ per minute of estimatedMinutes (minimum 5). Returns sessionId + connectUrl.

Brave — keyword web search over Brave's own independent index (not Google/Bing). Pass `q`; optional `count` (1-20), `offset` (pagination), `country` (2-letter), `search_lang`, `freshness` (pd|pw|pm|py or YYYY-MM-DDtoYYYY-MM-DD), `safesearch`, `result_filter` (comma list: web,news,videos,faq,discussions), `extra_snippets` (up to 5 alternative excerpts per hit). Returns ranked results with snippets — no page content; chain a scrape action for full text. The independent keyword-index alternative t…

Reveal the professional email behind a LinkedIn profile via Tomba. Pass `url` (a linkedin.com/in/… profile URL); `full: true` returns ALL associated emails. Alternative to prospeo:enrich when you have the LinkedIn URL. 1 Tomba credit.

List archived snapshots of a URL from the Internet Archive Wayback Machine (CDX index). Pass `url` + optional `from`/`to` (YYYY..YYYYMMDDhhmmss prefixes) and `limit`. Returns [timestamp, original, mimetype, statuscode, ...] rows — feed a row's timestamp to wayback/fetch. THE source for "what did this page/pricing/docs say in <year>" and for recovering dead links.

Linkup — AI web search returning a cited answer or ranked results. Pass `q`; optional `outputType` (sourcedAnswer default — synthesized answer with sources | searchResults — raw ranked hits | structured — pass `structuredOutputSchema` JSON-schema string), `fromDate`/`toDate` (ISO), `includeDomains`/`excludeDomains`, `includeImages`, `includeInlineCitations`. Standard depth — one-pass retrieval. For hard multi-hop questions escalate to linkup/deep-search (5¢).

CRW — Start an ASYNC multi-page crawl from a seed URL, following links. Pass `url`; optional `maxPages` (1-100, default 100 — vendor bills 1 credit/page so keep it tight), `maxDepth` (default 2), `formats`, `onlyMainContent`, `renderJs`. Returns `{ id }` immediately; KEEP POLLING `crw/crawl_status` with that id until status is completed/failed. One price covers the whole job.

ScrapingAnt — Scrape a HARD page through the 3M+ residential-proxy pool + headless Chrome: Cloudflare and anti-bot walls, geo-fenced content, sites that block datacenter IPs. Same params as scrapingant/scrape (`url` + browser/wait_for_selector/js_snippet/proxy_country/timeout/cookies). Returns `{ raw: "<html>…" }`. 4× the price of the datacenter rung — escalate here only after a plain scrape fails.

One XBRL financial concept for a PUBLIC company, all fiscal periods (data.sec.gov companyconcept). Pass `cik` + `concept` (us-gaap tag, e.g. RevenueFromContractWithCustomerExcludingAssessedTax, Revenues, NetIncomeLoss, Assets). Small structured response — the right way to answer "what was <public co>'s revenue", instead of scraping a multi-MB 10-K.

Verify an email address's deliverability via Prospeo (rides enrich-person with an email input — the 2026 API has no standalone verifier). Pass `email`. The response's `person.email.status` is the verdict (e.g. VERIFIED) with `verification_method`. 1 credit per found email; no vendor charge on a miss; results cached 90 days upstream.

TheirStack — List the BUYING-INTENT topics detected for a company from its job posts (each with confidence, mention count, first/last dates). Pass ONE of `company_domain` (preferred), `company_name`, `company_linkedin_url`; optional `keyword_slug_or` to check specific topics (slugs from theirstack/tech-catalog with keyword_type_not=technology), `confidence_or`, `limit`. The qualification signal before outreach: is this account showing intent for what you sell?

Google Trends (via x402atlas) — search-demand data: /trend, /interest-by-region, /related-queries, /related-topics. GET { endpoint, keyword, ... }. ~5¢ caps. Use for demand/market research alongside exa search.

TheirStack — Search companies by firmographics (industry, country, employee count, revenue, funding stage), tech stack (`company_technology_slug_or` — case-sensitive slugs from theirstack/tech-catalog), and hiring signals (`job_filters` + `min_num_jobs_found`: companies actively hiring for given roles/techs). REQUIRED: `limit` (1-5, vendor bills per company returned); `page` for more. Returns companies with the matching jobs/technologies attached — built for ABM, lookalike discovery, and compet…

Find companies whose recent JOB POSTS mention a technology via OpenFunnel (SYNC, 50 results/page): { tech } (e.g. "Snowflake") + optional `variations` (OR-matched phrasings, e.g. ["K8s"]), `lookback_days` (≤365), `page`, `enrich` (firmographics), `hq_country_code` (alpha-3), min/max_employee_count (filters auto-enable enrich), `include_evidence` (the job-post snippet — the outreach hook). Perfect for "sell to companies running X" and competitor-displacement plays.