Observe before execution
Shadow mode lets an operator inspect signals, rules, evidence, and decisions without allowing a trade to move forward.
About Balchemy
Balchemy helps a trading agent observe markets and make bounded decisions without giving it unrestricted access to a wallet. The operator can see the rule, evidence, mode, funding boundary, and result that shaped each consequential action.
The product thesis
Signals and market context enter through read-only research and source surfaces.
A plain-language rule and available evidence shape a bounded decision.
Mode, scope, risk, limits, and approval requirements determine whether execution is eligible.
Only the authorized path may submit, track, and reconcile a financial action.
What we refuse to blur
Calm design is useful only when it keeps consequence explicit. Unknown, degraded, blocked, submitted, and reconciled states are not cosmetic variants of success.
Shadow mode lets an operator inspect signals, rules, evidence, and decisions without allowing a trade to move forward.
An identity wallet proves who is signing in. The agent trades only from a separate funded boundary and inside active policy.
Runtime mode, scopes, limits, risk evidence, and approvals qualify the action before the execution layer receives it.
A proposal, submission, execution, and reconciled result remain distinct so the final state can be inspected honestly.
MCP, ERC-8004, walletless onboarding, replay protection, and chain-specific execution belong in the verification layer—not in the first sentence a trader has to decode.
Natural-language rules, public evidence, Hub operations, and commitment-aware onboarding.
Solana and Base paths, pre-trade policy, fees, callbacks, and reconciliation-safe lifecycle state.
Public product and privacy contact: burak@balchemy.ai.
Review a public record or read the control model before signing in.