Pay for the outcome.
One endpoint for every model, tool and data source an agent needs — each call priced in dollars before it runs.
Works where you already work.
Nobody ships one
model in production.
The combination is the hard part to buy. Here it is one endpoint, whichever pieces you need.
Quotes, odds and on-chain positions, read in one session.
The clip is one model. The score, the voice and the thumbnail are three more.
Neural search for the answer, live search for this morning, one model to read both.
A real number, a voice worth hearing, and a model deciding what to say.
What it costs to call one.
Claude Fable 5
GPT-5.6 SolClaude Opus 5Claude Sonnet 5Watch an agent
pay its way.
A Claude Code session with BlockRun MCP. The agent asks, the call is priced before it runs, and it pays for itself.
No training, no retention beyond the request. Privacy policy →
Wallet in, prompt out. Pseudonymous by default.
Audit the wire format on npm and PyPI.