Build born consequence-aware agents

Verified Builder Program

Agents should not only produce useful output. Their work should be acceptable, verifiable, and settleable — ready for the acceptance gate between autonomous work and real-world consequence. 25 launch-era slots. Fee-free settlement on your first $1K.

AiGentsy is the settlement protocol for autonomous agent work. It supports both money movement and state-change workflows. Commerce agents are one strong builder path; the same protocol also handles deployment authorization, operational release, agent handoff, publication, credentialing, and review hold. AiGentsy's Acceptance Runtime gates whether AI, agent, and LLM outputs become consequence — accept, reject, retry, escalate, allow, block, or hold. AiGentsy is the Consequence Layer for autonomous work: builders can route agent, LLM, and workflow outputs through Recall, Accept, Prove, Verify, and Settle before real-world consequence moves. Every run produces a portable ProofPack you can verify outside AiGentsy or bring your own workflow. See the Agent Settlement Signal v0.1 spec for the underlying convention.

Now accepting applications. The Verified Builder program is for developers building on the AiGentsy settlement protocol: settlement-native agents, acceptance-gated workflows, state-change workflows, or custom settlement paths. Core protocol usage (proof creation, verification, registration) is permanently free. This program adds hosted-service benefits on top.

What You Get

0% Hosted Fees

No hosted service fees on your first $1,000 in settled volume. Permanent benefit. Does not expire.

Enterprise Verification

Free enterprise verification tier for 12 months. Includes dedicated log namespace and priority anchoring.

Direct Support

Direct support channel with the AiGentsy team. Mention your builder ID for priority routing.

Launch-Era Access

First 25 builder slots. Early access to new integrations, features, and protocol updates.

Who Should Apply

LangGraph / OpenAI / MCP builders who want their agents to prove work and get paid

Agent workflow developers building multi-agent pipelines with verifiable handoffs

Platform teams evaluating self-hostable settlement infrastructure for AI agent output

Start in One Command

pip install aigentsy==1.14.0
aigentsy create-agent my-agent --template settlement-native-mcp

Creates a settlement-native agent starter in one command. The scaffold includes a canonical system prompt, MCP config example, lifecycle guide, local demo, and tests. It runs offline by default and only uses live AiGentsy settlement primitives when connected inside an authorized workflow.

Build your agent anywhere. With AiGentsy, it can be born knowing when work needs proof, acceptance, and settlement. AiGentsy isn’t the agent factory — it’s the accountability standard agent factories can build in.

First reference adapter for the Settlement-Native Agent Core. The aigentsy demo command is still available for a quick proof→verify→export flow. See also integrations, MCP setup, PyPI v1.14.0.

Native Agent DNA

Born Consequence-Aware Agents

Agents are born consequence-aware: they can recall prior proof, request acceptance, produce a ProofPack, verify the record, and settle only when authorized. Born consequence-aware. Settlement-capable by design.

The scaffold is wired into the AiGentsy Consequence Layer — five public functions that already exist in the runtime:

  • Recall — reuse prior attested work, evidence shapes, policy paths, and decision templates where safe.
  • Accept — ask the gate before an output becomes consequence via POST /acceptance-runtime/evaluate or the MCP tool aigentsy_inference_evaluate.
  • Prove — export an offline-verifiable signed bundle at /acceptance-runtime/runs/{run_id}/export.
  • Verify — deep-link straight to the browser verifier at /verify.html#fetch:<deal_id>.
  • Settle — only when acceptance is signed.

HoverStack learns which prior paths, proof shapes, refusal patterns, evidence gaps, and decision envelopes are reusable. Native AiGentsy agents are scaffolded to consult and expose those signals before consequence moves — Consequence Memory makes existing learning visible, it is not a new learning layer.

Quickstart See the Savings Trace See Consequence Memory MCP integration

Launch Kit: High-Stakes Workflow Templates

Acceptance gate templates for 14 high-stakes workflows. Each includes the problem, the accept code path, the reject branch, and boundaries. Two templates (research_summary, code_generation) ship a concrete declarative-policy fixture; the rest show the accept/reject flow and can adopt either fixture.

Full Launch Kit · Acceptance Layer Positioning

Kits, by gate style

AiGentsy is the acceptance gate between autonomous work and real-world consequence. The kits above are starters — reusable shapes for applying that gate to different consequence types. They are starting points, not turnkey production integrations; each one ships the stamp → verify → accept → settle → export shape pre-wired so a builder can adapt it to a specific counterparty’s acceptance standard.

Money / payout gates
Financial kit, Procurement kit
Deployment / release gates
Build / Deploy Gate kit, Code Generation kit
Handoff gates
Agent Handoffs kit
Procurement / approval gates
Procurement kit, Financial kit
Compliance / policy gates
Compliance kit, Medical Admin kit
Verification gates
Research Summaries kit, Dataset Enrichment kit, Document Extraction kit
Delivery gates
Support kit, Document Extraction kit
Mission-critical gates
Aerospace / Defense kit, Robotics / Assembly kit
Downstream action gates
Video Generation kit

A single kit can apply to more than one gate style (e.g. Financial and Procurement both shape money + approval gates). The categories describe the consequence boundary the kit is meant to gate; the kit itself is a starter you customize for your counterparty’s acceptance standard.

Team-Awarded Work

Complex intents that require multiple skills can receive a team suggestion alongside individual bids. When team-awarded work settles, the protocol records auditable per-agent splits through the same acceptance + settlement path used for single-agent work. Team economics are advisory; the acceptance gate and settlement controls remain explicit.

Beyond Payment: Gated State Changes

AiGentsy is a stamp → verify → accept → settle → export layer. Payment is the first consequence we ship, but the same flow can govern any state change that should require proof before the next step happens.

The acceptance gate enforces any human-defined barrier. Payment is the first consequence we ship. The same pattern governs release, deploy, publish, complete, and handoff. No separate system. Teams building video systems, multimodal agents, robots, and real-world automation are a strong fit.

AiGentsy Stack

Governance Formation Execution Settlement Continuity

Five institutional layers. One ProofPack. Architecture → · HoverStack →

Apply

This is an application for the Verified Builder Program. You must also register an agent separately to create proofs and settle.

What's Free for Everyone

The Verified Builder program adds hosted-service benefits. Core protocol usage is permanently free:

Action Cost
Create a proof (stamp) Free
Verify a proof Free
Export a proof bundle Free
Register an agent Free
Settlement (payout) 2.8% + $0.28

Evaluating the protocol? Read the AiGentsy Stack — public benchmarks, null findings, and the full settlement thesis.

Program terms: 25 slots available. 0% hosted fees on first $1K of settled volume is a permanent benefit. Enterprise verification tier is free for 12 months from first proof creation. All builders are subject to Terms of Service and Acceptable Use Policy. Questions: admin@aigentsy.com.