An AI workspace you can run yourself. Every answer shows its work.
One question becomes a plan. A team of agents runs it. Every step keeps its evidence.
Go gateway. Python brain. Rust core. Yours to host.Which vendor actually met the SLA last quarter?
- contracts
- invoices
- uptime log
- support tickets
- MSA §4.2
- uptime export
- ticket audit
Verity exists to make good AI cheaper for everyone, simpler for anyone, lighter on the planet, and sharper than the tool you use today.
AI for all
SpiderMesh runs the work across ordinary machines people already own, so the price follows spare capacity instead of rented racks.
How SpiderMesh worksAI, simplified
Plain words in, finished work out. Nothing was removed to make it simple.
AI for nature
A datacentre drinks water to stay cool. SpiderMesh borrows machines that are already running and already cooled, so the work adds no new cooling load.
How SpiderMesh worksSmarter AI
Two nodes have to agree before a result counts, and every answer arrives with the evidence behind it.
The protection is built into the way the work moves.
- L1At the edgeUnwanted input is stopped before it reaches your work.
- L2On the way inEach request is checked before it moves forward.
- L3Your spaceYour work stays tied to your workspace.
- L4In the answerOutside text cannot quietly change the job.
- L5In your dataThe system closes the door when it cannot prove where something belongs.
One command, your hardware.
docker compose -f infra/docker/compose.yaml up -dPostgres, Redis, Qdrant and the three Verity services.
Start with one clear next step.
Ask a question, watch the work take shape, and open the details when you want to look closer.