Live timeline, on-chain audit trail, and custom alerts for any AI agent. Connect via npm, npx, or Skills.
Wire any agent in once. Every prompt, tool call, edit, and decision streams to Voight automatically. Searchable, replayable, audit-ready.
See decisions as they happen — not post-mortem, three hours after the liquidation. A heart monitor, not an autopsy.
Every log is hashed and anchored on Solana. Tamper-evident by architecture — not policy. Regulators accept it because the chain is the witness.
We learn each agent's baseline — tool mix, latency distribution, tx patterns. You get alerted the moment behaviour drifts. No rules to write.
A public, searchable index of every agent in the on-chain Agent Registry — plus the full x402 payment graph across chains. Updated as blocks roll in. No login. No paywall.
No migration required. Voight sits next to your existing agent code and captures what matters.











You didn't start a company to write evidence logs. The chain handles it. You export.
Pause, intervene, and acknowledge anomalies from the dashboard. Every oversight action is logged alongside the agent's own record.
Each event is hashed into a Merkle root anchored on Solana every slot. Integrity is verifiable without trusting Voight.
Generate signed audit packages by date range, agent, or incident — PDF + JSON + chain proofs — with one click.
No per-event surprises. No enterprise gating of the basics.
For solo devs and prototypes.
Built for teams ready to scale.
For regulated deployments and prop desks.
Free forever for solo devs. No credit card. Three lines of code, then a dashboard.