Cerebras

Native

AI · United States · Website ↗

Connect your Cerebras account to power Aria with wafer-scale inference — GPT OSS 120B at ~3,000 tokens/second, the fastest way to run open-weight models.

Governed by default

Like every connector, Cerebras runs inside the ApiSpi governance layer: admins can set it to full, read-only, or approval-required access, every call is audit-logged and firewall-screened, and spend caps apply.

Put Cerebras to Work

Connect it in minutes and let your agents act — inside guardrails you control.

Start Free Week Talk to Us