I built @genailab_ the open-source terminal agent in one native binary. Minimal overhead. Maximum agent.
CA : 0xb95679076e952aa62ab87f6c1be75062f6d61820
$SAN is now live @ponsdotfamily
CA: 0xb95679076e952aa62ab87f6c1be75062f6d61820
San the open-source terminal agent in one native binary now has a token on Pons.
One binary. Any model. Now on Robinhood Chain.
https://t.co/s2QGlDGxqx
I switch models without switching tools.
DeepSeek for one task, Ollama local for the next San lets me pick the right model per step.
One binary. Any model.
https://t.co/s2QGlDGxqx
San starts in 0.01s. Downloads in 12 MB. Runs on 2.3k tokens of harness.
That's why I run it every day and why I launched $SAN.
Small footprint. Maximum agent.
Beware of clones.
A lot of fake $SAN tokens are popping up. This is the only official one:
CA: 0xb95679076e952aa62ab87f6c1be75062f6d61820
Robinhood Chain. Verify the CA before you buy.
I use San because there's less framework and more room to work.
One binary. No Node, no Python runtime to babysit. Just the task, and the agent that does it.
I launched $SAN because San is the agent I actually run every day.
One binary. 12 MB. 0.01s startup. No bloat, no lock-in.
The token is for the people who build with it.
One binary. Everywhere.
San runs on macOS, Linux, and Windows β plus Pi, Jetson, servers, and CI. Even offline with a local model.
One binary. Any model.
https://t.co/s2QGlDGxqx
$SAN is launching on @ponsdotfamily ,on Robinhood Chain.
San is the open-source terminal agent runtime β one native binary, no Node.js or Python runtime. 12 MB download. 0.01s startup.
https://t.co/s2QGlDGxqx