CleverThis is here

Today we’re launching CleverThis — the agent command center for everyone shipping with LLMs and everyone using agents to get real work done. One HTTP endpoint speaks the OpenAI API and routes across 130+ providers and thousands of models from the open models.dev catalogue. Point any OpenAI SDK at it, swap the API key for a CleverThis token, and start sending traffic. No code changes, no provider lock-in, no separate billing relationships.

Underneath, every request carries a frozen pricing snapshot, an explicit provider preference chain, and a full provenance trail. If your first-choice provider degrades, the gateway falls over to the next candidate inside the same streaming response. Failed attempts aren’t billed.

On top of the raw gateway sits the agent runtime. Hierarchical agents and hierarchical actors decompose hour-long, multi-step jobs into isolated subtasks with their own context windows — so the work that flattens single-shot agents actually finishes. You describe a prompt + tool surface in YAML (the open CleverAgents Actor format), CleverThis compiles it to an immutable content-addressed identifier, and you publish it under a stable name in a namespace. Customers call it like any other model id; you can roll versions forward or back without touching their integration.

The control plane speaks A2A JSON-RPC: every operation the CleverThis webapp performs is a public surface, so any A2A-compliant client, the CleverAgents CLI, or your own HTTP client can drive the platform with full parity. No private endpoints, no lock-in to our UI.

CleverAgents — the MIT-licensed CLI counterpart at git.cleverthis.com/CleverAgents/CleverAgents-core — mirrors much of the CleverThis runtime locally. Develop and run agents on your own hardware, then push the same actor bundles up to CleverThis when you need scale, observability, and managed infrastructure. The artifacts move unchanged.

Sign up to try the sandbox; the data plane is live, the Endpoints browse + detail pages are wired in, and the wallet + BYOK flows are ready for early users. Follow this feed for product and engineering updates as the platform fills out.