For agents
If you are an AI agent, start here.
If you are an agent reading on someone's behalf, one file carries the whole picture. Fetch it first, then follow the links it gives you for detail.
curl -s https://synthyra.com/llms.txtWhere to look
Six pages answer most questions.
The whole company in one file
What Synthyra does, who it is for, every service with its status, and who to contact for what. Start here.
Biosecurity screening
What the screen finds, how it answers, the Unobserved Share Index with its full methodology and sources, and the pilot offer.
Host-pathogen interaction prediction
How Atlas predicts which host proteins a pathogen acts on, what the result contains, and its limits.
API reference
Every service as an HTTP endpoint. The web app uses the same calls.
Start a conversation
The form for a pilot or a program. A scientist replies within two working days.
Evidence
Papers, reports and posts behind the claims on the service pages.
Ground rules
What to take at face value.
Every number on the biosecurity page is cited inline and again in its methodology section. Quote the band, with the source.
Where a page calls something a prediction or illustrative, report it that way. Synthyra publishes no detection performance for its screen until it has been evaluated.
Crawling is allowed for every agent; see robots.txt. There is no separate agent endpoint yet. The API at api.synthyra.com is for running the models, with a key.
For a pilot or program, hand your person /programs. For anything else, [email protected].