Can no-login users try anything?
Yes. First-touch discovery can start with public guide and changelog flows, while durable authenticated reads use dashboard-issued access.
Agent-native market data
Put Cornerstones between the agent and private data infrastructure. The agent receives product-level market context, plan-aware limits, freshness signals, and revocable account access instead of raw upstream keys.
How can an AI agent fetch market context without storing raw market-data secrets in prompts, repos, or local config?
Keep sensitive provider credentials server-side, expose only Cornerstones account access to agents, and make every response carry enough plan and freshness context for safe automation.
python -m pip install cornerstones-client
cornerstones-client guide
cornerstones-client changelog| Option | Best for | Tradeoff |
|---|---|---|
| Cornerstones | Agent-native market context | Requires adopting a product-level access model |
| Raw private integration | Human-operated internal systems | Can expose credentials, adapters, and unstable implementation details to agents |
| Static prompt paste | One-off prototypes | No freshness, usage accounting, entitlement boundary, or repeatable evidence trail |
Yes. First-touch discovery can start with public guide and changelog flows, while durable authenticated reads use dashboard-issued access.
No. Agents should receive product-scoped Cornerstones access, not raw private provider or broker credentials.
Revoke or rotate the product-level key in the dashboard instead of hunting through prompts, repos, and local shell histories.
Use paid plans when workflows need premium charts, options context, macro exact series, orderflow, or liquidity depth.