Get paid to build the AI frontier
Scoped, treasury-funded work that strengthens the Claude ecosystem — paid in USD value plus $C2E. Here's a preview of what's coming.
Preview. The board opens at launch — these are examples of the kind of work the treasury will fund across 8 categories. Rewards shown are indicative.
Autonomous research agent with Claude tool-use
A multi-step agent that plans, searches, and synthesizes cited answers using Claude's tool-use API.
Edge inference proxy with smart caching
A latency-optimized gateway that caches semantically-similar Claude requests at the edge.
Claude-powered code review extension
An inline PR reviewer that streams structured suggestions and risk flags directly in the editor.
Interpretability view for agent traces
Make agent behavior legible — decision steps, tool calls, and token economics in one view.
Open eval harness for agent reliability
A reproducible benchmark measuring tool-call accuracy, latency, and cost across agent frameworks.
CLI scaffolder for Claude agent projects
A generator with batteries-included tracing, evals, and deploy targets out of the box.