# FundingData
Agent-callable global startup funding data. Every record links its source: the published article, or for older records the tech.eu deal-database entry. What each year contains: /coverage.
ACCESS: data endpoints require an API key (X-Api-Key: fdk_… or Authorization: Bearer fdk_…) or a member session; without one they return 401.
Keys: https://fundingdata.com/app#/account · plans and limits: https://fundingdata.com/pricing · coverage: https://fundingdata.com/coverage
MCP: POST https://fundingdata.com/mcp (Streamable HTTP, same key).
Filter by region/country and by founder origin (e.g. Turkish-founded companies worldwide).

## Plans (USD, from lib/plans.js)
- Free: $0 · last 12 months of rounds · 1,000 records/month via API and MCP (key required) · attribution to FundingData required · no AI answers (ask / semantic_search)
- Pro: $99/month (2,500 records/month) or $790/year (5,000 records/month) · all years · CSV export · alerts
- API: $2,400/year · 25,000 records/month · bulk feed + incremental sync (/v1/rounds/changes) · webhooks
- Data licences (file/feed, display, redistribution): from $12,000/year — https://fundingdata.com/pricing

## Markdown twins (text/markdown)
- /company/<name>.md — one company's funding history. Ex: /company/Brix.md
- /investor/<name>.md — one investor's deals + portfolio. Ex: /investor/Sequoia%20Capital.md
- /rounds.md?<filters> — filtered rounds table, ≤200 rows, same filters as /api/rounds. Ex: /rounds.md?geo=Europe&period=2026&limit=20
Bare /company/<name> · /investor/<name>: HTML+JSON-LD for text/html Accepts (browsers, crawlers), the markdown twin otherwise.

## Tools (remote MCP: POST /mcp, JSON-RPC 2.0)
- funding_rounds(sector, stage, geo, period, investor, founderOrigin, turkishFounded, market, investorCountry, minAmountUSD, limit) — List global startup funding rounds. geo accepts a region (Türkiye, Europe, MENA, Americas, Asia, Africa) or an HQ country code (TR, US, SA, DE, IN, …). founderOrigin filters by founder origin regardless of HQ (e.g. TR = Turkish-founded worldwide). turkishFounded=true is shorthand for founderOrigin=TR. market=XX (ISO2) = the whole market of a country: HQ there OR founders from there OR at least one investor headquartered there; each row then carries marketReasons [hq, founder, investor]. investorCountry=XX = at least one investor headquartered there. Other filters: sector, stage, period (YYYY or YYYY-Qn), investor, minAmountUSD, limit. Each result carries a source citation.
- company(name | domain) — Funding history of one company — HQ, founders, total raised, rounds, investors, citations. Look it up by name or by website domain.
- investor(name) — Activity of one investor — deals, capital deployed, sector/geo mix, portfolio.
- ecosystem_stats(geo, period, sector, stage, founderOrigin, turkishFounded, market, investorCountry) — Aggregate stats for a geo/period (optionally founderOrigin / turkishFounded, market=XX or investorCountry=XX as in funding_rounds) — total, by sector, by quarter, by stage, by country, by region. Pass sector=Travel for one sector (its countries, stages, top rounds and investors).
- fund_announcements(geo, limit) — New investment FUND announcements (a VC/PE firm raising/closing/launching a fund — distinct from a startup funding round): firm, fund name, size, vintage, close type, strategy, HQ geo. Each carries a source citation. Filters: geo (ISO2 country), limit.
- tech_news(category, geo, limit) — Notable startup / tech-ecosystem news beyond funding (launches, expansions, partnerships, products, people moves, market entries, regulation, milestones) at named tech companies/VCs — title, summary, mentioned companies/people/funds, sector, geo, with a source citation. Filters: category, geo, limit.
- likely_raising(region, sector, limit) — Companies likely to raise their next round soon — a transparent, citable ESTIMATE (not a prediction): ranks companies whose last disclosed round was 9-33 months ago (peak ~18mo), weighted by stage and round count. Returns a methodology note. Filters: region, sector, limit.
- semantic_search(q, kind, limit) — Natural-language semantic search over the cited content corpus (notable tech-ecosystem news + events: funds, M&A, IPOs). Finds items by MEANING, not keywords — e.g. "AI for drug discovery", "climate hardware", "fintech in Africa". Returns ranked hits with a relevance score (0–1) and a source citation. Filters: q (required), kind (news|event), limit.
- regulatory_actions(geo, limit) — Regulator actions affecting the tech / fintech / startup ecosystem — licence grants & revocations, capital-markets and competition (M&A) decisions, rules. Coverage: Türkiye (BDDK, Rekabet Kurumu, Official Gazette) plus UK, EU, US and India regulators. Each carries a source citation. Filters: geo (ISO2), limit.
- dossier(name, kind) — Unified entity dossier for one company, investor, or person — funding history + fund/M&A/IPO/regulatory events the entity is named in + news mentions + authoritative GLEIF legal name/LEI/HQ country, every item cited. One call, the whole graph around an entity.
- entity_graph(name, kind, edgeType, limit) — Relationship graph around one entity — the actual typed, weighted edges, each with a source citation. For an investor: its co-investors (who it syndicates with) and its portfolio. For a company: its investors, who acquired it, and its founders. For a person: the companies they founded. Edge types: co_invested, invested_in, acquired, founded_by. Filters: name (required), kind (company|investor|person, default investor), edgeType (optional, one of the four), limit.
- top_investors(limit) — The most influential investors ("smart money") ranked by PageRank centrality over the co-investment graph — i.e. who syndicates with the most well-connected investors. A transparent, derived deal-sourcing signal. Each carries its co-investor breadth and network volume. Filter: limit.
- emerging_sectors(region, windowDays, limit) — Early-trend detection: which sectors are ACCELERATING — ranked by funding velocity + acceleration (2nd derivative) while still a small share of total volume (the weak-signal pattern that gives lead time before a theme is obvious). Each sector carries its velocity, acceleration, share, and recent CITED rounds. A "what is heating up" signal. Filters: region (Türkiye|Europe|MENA|Americas|Asia|Africa), windowDays (default 60), limit.
- accelerators((none)) — The accelerator / demo-day programs tracked (Y Combinator, 500 Global, SOSV) — company count, batch count, and how many of each cohort have since raised. The startup-SCOUTING layer: companies surfaced BEFORE they raise.
- accelerator_cohort(slug, batch, q, limit) — The startups selected into one accelerator program, optionally one batch — each with its one-liner, batch, and funding raised since (if any). Cited to the program's own roster. Filters: slug (required, e.g. yc | 500global | sosv), batch (e.g. "Winter 2026"), q (free-text over name/one-liner), limit.
- sec_filings(name) — SEC filings of one company: the SEC registrant(s) FundingData linked to it from its own EDGAR filings (S-1/F-1/424B4, Form D/C/1-A) — legal name, CIK, tickers, SIC — and the latest filings (S-1, DRS, prospectus, Form D, 8-K, 10-K…) with direct links to each document. Use it to cite primary documents.
- fund_marks(name) — Mutual funds' fair-value marks on one PRIVATE company, from SEC N-PORT filings (Fidelity, T. Rowe Price, Franklin, …): per quarter-end period and share class (Series G, Common, …), how many funds hold it, the median / min / max US$ per share and the total value held, each with the filing as citation. The only regularly published independent price marks on private companies; filings appear ~60 days after the period.
- unreported_filings(days, min, q, limit, offset) — SEC Form D placements of tech and biotech companies that have NO funding round in the record yet: the issuer told the SEC it sold securities, and no press-reported round matches (same legal name, first sale -21/+120 days). Filing facts with the SEC link as citation, not estimates. Newest first; filter by window (days), minimum amount sold (min, US$) and issuer name (q). Venture-fund vehicles and listed issuers are excluded. Paid plans.
- report_correction(recordId, recordKind (round|event|company|investor), field, proposedValue, evidenceUrl, note) — Report a wrong field on a record (a round or event id, or a company / investor name) with the proposed value and an evidence URL. Stored for review by the FundingData team; nothing changes in the data until a reviewer applies it. Needs an API key or a signed-in account; 20 reports per account per 24 h.
- resolve(name, kind (company|investor) | domain) — Resolve a free-text name to the exact company / investor names the other tools accept — exact matches flagged, close names listed, best set only when one candidate clearly wins. Call it first when the name came from a user or another model. A website domain (domain=) resolves to the company recorded with it.
Ex: curl -X POST https://fundingdata.com/mcp -H 'Content-Type: application/json' -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"company","arguments":{"name":"Brix"}}}'

## REST (JSON)
- /api/rounds?sector=&stage=&geo=&period=&investor=&founderOrigin=&turkishFounded=&market=&investorCountry=&minAmountUSD=&limit=&offset= — Ex: /api/rounds?sector=AI&period=2026&limit=5
- /api/company?name= · /api/investor?name= — Ex: /api/company?name=Brix
- Versioned: /v1/rounds, /v1/company, /v1/investor, /v1/companies, /v1/investors, /v1/search, /v1/stats, /v1/facets, /v1/events, /v1/news (real 404s) · spec: /openapi.json
Auth for API/MCP clients: X-Api-Key: fdk_… or Authorization: Bearer fdk_…
- /api/stats?geo=&period=&g=month|quarter|year (alias /api/report) — aggregates. Ex: /api/stats?geo=Europe&period=2026
- /api/companies?q=&region=&sector=&sort= · /api/investors?q=&sort= · /api/search?q=
- /api/facets (filter enums) · /feed.json (bulk: most recent 5000 rounds)

## Datasets (aggregate JSON snapshots · authorized users, cite FundingData · refreshed ≤10 min)
- /datasets — index
- /datasets/rounds-by-year.json · /datasets/rounds-by-country.json · /datasets/rounds-by-sector-year.json · /datasets/investors-top.json

geo accepts a region (Türkiye, Europe, MENA, Americas, Asia, Africa) or an HQ country code (TR, US, SA, DE, IN, …).
amountUSD null = undisclosed (listed, never summed). amountEUR = the same amount at the ECB rate on the round date (eurRate, eurRateDate). Every record includes a source citation URL.
stageEstimated:true (tech.eu deal-database rows) = the source states no stage and the stored stage is a size bucket; customer keys get stage null + stageEstimate, ?stages=honest|stored chooses.
License: data is licensed to authorized users with attribution to FundingData; bulk redistribution needs a written agreement (/terms).
Full endpoint + parameter reference: /llms-full.txt · HTML reference with field definitions: /docs · Methodology: /methodology