samson://jamestate: building

Compute and research for AI agents.

Samson is the compute and research arm of Jame. Agents need somewhere to run and models that suit the work. We are building both, in the open.

See the architecture GitHub
01 / tracks03 listed

What Samson is building.

IDTrackWhatState
01Agent ComputePay-as-you-go compute for agents running on Jame. No plans and no seats: an agent pays for what it uses.░░ building
02Samson ResearchOpen research on agent infrastructure: how autonomous agents execute, coordinate, and which models carry the work best.░░ building
03Local ComputeAgents run on your own hardware through JameClaw.██ live
02 / runs00 public

Runs.

> no public runs yet. the first one will be listed here.
03 / where it fitslayer 03 of 03

The third layer of Jame.

┌─────────┬──────────────────────────────┐
│ AGENTS  │ JameClaw · Clawie · JameFlow │
├─────────┼──────────────────────────────┤
│ MODELS  │ agentic model · model board  │
├─────────┼──────────────────────────────┤
│ COMPUTE │ Samson  ◄ you are here       │
└─────────┴──────────────────────────────┘

Jame stays free and open source. Samson is the only place where usage is paid for.

Earlier Samson experiments, including ValOs, remain online as an archive at samsonresearch.xyz.