five agents. the same tests, five times.
every agent runs the whole test suite on its own.
and all of it squeezes through one bottleneck: your ci.
so every change waits in line behind the others.
a local-first developer tool
antil runs a separate app for each worker, shares heavy resources, and reuses test results only while the code and environment they tested still match.
free and open source on your machine · paid cloud across machines · early access
every agent runs the whole test suite on its own.
and all of it squeezes through one bottleneck: your ci.
so every change waits in line behind the others.
you, and every agent you start.
its own app on its own address, ready only when it really answers.
your agents open it in their own browser. nothing to integrate.
a migration on a shared database would hit everyone.
so the worker that changes the schema digs its own chamber first.
it brings its own copy of the data. everyone else keeps working.
somebody has to run the tests first.
antil records exactly what they proved: the code, and the environment it ran against.
where one ant has been, the next one doesn’t have to go.
your change touched one tunnel.
only that part gets walked again.
anything antil can’t map always runs. nothing is skipped on a guess.
a teammate’s proof arrives from the next nest.
ci comes up to check it: who produced it, and whether its inputs still match.
if it holds, ci takes the seal instead of running everything again.
on your laptop: you and your agents.
across your team: your colleagues’ nests.
shared environments and ci: the whole field. idle resources come home instead of billing for weeks.

field note
on your machine, antil is free and open source, for you and every agent.
sharing proof, trust or resources across people and machines is antil cloud.
paid, and in early access.
the first wave is the local tool on macos. we’ll write when yours is ready.
