IARPA DECIPHER

IARPA DECIPHER

Funds United States teams building multilingual language tools for difficult intelligence vocabulary and specialised terminology.

DECIPHER is a research program administered by the Intelligence Advanced Research Projects Activity (IARPA), a U.S. agency within the Office of the Director of National Intelligence (ODNI). The program operates under IARPA's Emerging Technology Accelerator (ETA) framework and targets a specific failure mode in intelligence analysis: the inability of standard translation and natural language processing tools to handle specialized, obfuscated, or emergent language. DECIPHER funds development of capabilities to detect and define three classes of problem language in multilingual text collections — expressions that are hard to disambiguate (such as domain-specific acronyms), language used to deliberately obscure meaning (coded language), and emergent jargon or slang for which accurate translations do not yet exist. Awards are issued as Prototype Other Transaction Agreements under solicitation DOI-ETA-FY26-30, administered by the U.S. Department of the Interior on behalf of IARPA.

Eligible prime contractors must be U.S.-based entities, including companies, universities, nonprofits, and non-traditional defense contractors. Non-U.S. organizations may participate only as sub-contractors to a U.S.-prime team. Government agencies, FFRDCs, and UARCs cannot serve as primes, and classified proposals are not accepted. OTA qualification requires non-traditional participation, small-business exclusivity, or at least one-third non-federal cost-sharing. The ETA framework typically produces awards in the $1 million to $5 million range for performance periods of 12–24 months, though DECIPHER-specific award amounts are not published.

The solicitation follows a white-paper-first process: proposers submit a white paper, and only teams invited after screening may submit a full proposal. White papers for the initial wave were due June 30, 2026, with continuous rolling submissions accepted thereafter until the program receives sufficient responses and closes. Evaluation emphasizes technical innovation, feasibility, and team capabilities. Research teams combining multilingual NLP, corpus linguistics, and low-resource language expertise are the most competitive applicants.