@pipeworx/pharma-deals

Connect: https://gateway.pipeworx.io/pharma-deals/mcp · Install: one-click buttons

Tools: 3

Worldwide biopharma licensing and M&A with the deal economics — upfront payments, milestone totals, royalty terms, acquirer, target and price — read out of the SEC 8-Ks the parties file when they announce.

Tools

• pharma_licensing_deals({since?, until?, query?, company?, therapy_area?, min_upfront_usd?, limit?}) — worldwide licensing, collaboration and partnership deals. Returns company, counterparty, asset, territory, upfront, milestone total, royalty language and therapy area, each with the sentence it was read from. Answers “recent pharma licensing deals”, “biotech out-licensing this quarter”, “who partnered on an ADC”, “which deals had an upfront over $100M”.
• pharma_deal_terms({accession, cik?}) — full economics for ONE announcement: upfront, near-term and total milestones, royalty tiers, equity component, territory and option rights, each with its source quote. Answers “what were the terms of that deal”.
• pharma_ma_deals({since?, until?, query?, include_unpriced?, limit?}) — pharma and biotech acquisitions with acquirer, target, per-share price, transaction value and any contingent value right, joined to the FTC Hart-Scott-Rodino early-termination notice for the same party where one exists. Answers “recent biotech acquisitions”, “what did that acquisition pay per share”, “did antitrust clearance come through”.

Auth

Keyless for the SEC half — EDGAR full-text search needs no credential.

pharma_ma_deals additionally calls the FTC HSR notice API, which uses an api.data.gov key. The gateway supplies the shared key automatically; a direct caller can pass their own as _apiKey (free at https://api.data.gov/signup/). If that lookup is unavailable the tool still returns the SEC half and says the clearance join degraded — it never reports a failed lookup as “no clearance on record”.

Coverage — read this before trusting an absence

Deals between two non-SEC-filing parties are invisible here. The source is the 8-K a party is required to file, so a deal between two private companies, or two non-US-listed ones, is announced nowhere in this data and will simply not appear. Every response repeats this in coverage_note.

Two further limits:

• Results are screened to filers with a drug-company SIC (2833, 2834, 2835, 2836, 8731). That is what keeps non-pharma licensing — franchising, software, trademark — out of the results. Each response reports how many filings the screen removed.
• pharma_ma_deals defaults to deal announcements, which state a cash price per share. Amendments, completions and all-stock mergers are excluded unless you pass include_unpriced: true; they name the parties but usually state no consideration.

How the numbers are read

Every amount is extracted from the announcement’s own prose and returned beside the sentence it came from, so it can be checked rather than trusted. Three rules the extraction follows, each of which exists because the naive version produced a confident wrong number on a real filing:

• The upfront is the money adjacent to the phrase, not the first dollar sign in the sentence — a release that opens with the company’s cash balance would otherwise report that balance as the upfront.
• The milestone total is read from its own cue (“up to”, “in the aggregate”, or money adjacent to the word “milestone”), because the milestone sentence routinely restates the upfront first.
• A term that is absent was not stated in the filing. It is not a zero. Companies frequently announce “an upfront payment” without disclosing its size; the amount comes back null and the quote shows why.

counterparty, asset and therapy_area are best-effort readings of the same text. therapy_area is decided by which area’s vocabulary appears most often, and therapy_area_mentions says how thin the call was.

For M&A, acquirer and target come from the filing’s own defined terms (“Parent”, “the Company”). When a filing does not use that convention they are left null, parties lists every named entity with its defined term, and roles_basis says why — because a filer is the acquirer about as often as it is the target, and guessing inverts the deal.

Data sources

• SEC EDGAR full-text search — https://efts.sec.gov/LATEST/search-index
• SEC EDGAR filing archives — https://www.sec.gov/Archives/edgar/data/
• FTC Hart-Scott-Rodino early-termination notices, via api.data.gov — https://api.ftc.gov/

The deal vocabulary and money parser are shared with @pipeworx/china-pharma, which asks the same question scoped to China.

Tools

• pharma_licensing_deals — Worldwide biopharma licensing, collaboration and partnership deals with their economics — the two companies, the drug or asset, the territory, the upfront payment, the milestone total and the royalty
• pharma_deal_terms — Full deal economics for ONE biopharma announcement, given its SEC accession number — the upfront payment, near-term and total milestone amounts, royalty tier wording, any equity component, the license
• pharma_ma_deals — Pharmaceutical and biotech mergers and acquisitions — acquirer, target, the per-share and total transaction value, and any contingent value right — taken from Agreement-and-Plan-of-Merger 8-Ks filed b