← zentient.studio

The first cell of the
Sentient Habitats

AI co-steward for regenerative coordination. Turns regenerative public-goods ideas into verifiable action plans — with claim discipline, evidence requirements, partner routing, and a 7-day execution path.

Apache 2.0 engine + demo dataset Commercial license: full Ecoterium substrate Claude Sonnet 4.6 + GLM-5.1 OpenClaw-aligned

Regeneration is not repair.
It is the active design of living systems.

Scale 1 · Planetary

The Sentient Habitat

A planetary-scale biological operating system. Not a building. A coordination layer for living infrastructure.

Scale 2 · Urban

The Living City

Re-engineering the urban grid into a self-sustaining biome. Energy, water, food, intelligence — woven, not bolted on.

Scale 3 · Replicable Unit

Zentient Regen — the first cell

Every planetary biome begins with a single, replicable unit. Zentient Regen is that cell — open-source, copyable, alive.

Scale 4 · Now

The Ground Reality

Transforming empty urban lots into high-yield food forests. The immediate execution layer. Not aspirational. Already deploying.

Scale 5 · Engine

The Invisible Engine

Raw data → AI worker swarm → structured action. Satellite, soil, grid feeds become precise mapping, resource allocation, localized growth.

Claim Discipline · Rigor over Rhetoric

We generate cryptographic proof plans, verify system state changes, and track biological execution. We do not certify carbon credits, trade impact derivatives, or make overblown climate claims.

Safety Doctrine

Zentient Regen does not certify impact. It creates a proof plan: what must be verified before a regenerative claim should be trusted.

Six structured outputs

You describe a regenerative opportunity in plain language — "let's turn an empty lot into a community food forest with solar, compost, and youth programming." The agent returns six structured outputs. Every claim it produces is labeled. Every step it routes is auditable.

OUTPUT 01
Classification
What kind of project is this? Real estate / food / energy / waste / community / funding / partnership — mapped to the regenerative-action database.
OUTPUT 02
Impact Map
What outcomes does it claim or could plausibly produce? Each linked to Drawdown, SDG, or peer-reviewed source.
OUTPUT 03
Proof Plan
What evidence is needed before any claim can be trusted? Lot-readiness, stakeholder consent, funding-readiness, and lifetime impact protocols.
OUTPUT 04
Partner Routing
Which human, organization, or other agent should act next? City planning, fiscal sponsors, schools, cooperative extension, peer bots in the swarm.
OUTPUT 05
7-Day Action Plan
Concrete, falsifiable next steps. Day-by-day deliverables with owners. The bot drafts and routes; humans decide.
OUTPUT 06
Claim Discipline
Every claim labeled proven plausible speculative not yet verifiable

Generate a proof plan in 30 seconds

Describe a regenerative opportunity. Zentient Regen returns a 9-section proof plan with impact map, monetization & Maslow-needs match, evidence checklist, funding fit, a 7-day hire-flow action plan, and route-to-swarm. Reasoning runs on Claude Sonnet 4.6; the regional precedent scan calls Z.AI GLM-5.1 (gracefully stubs offline).

Live demo · Powered by zentient.studio/api/regen/evaluate

Generation takes ~20–30 seconds — the agent loads the action library, reasons through 9 sections, and runs the regional scan in parallel. Output is plain markdown; claim labels color-code automatically.

Output will render here once you generate a proof plan.
Or run it locally
CLI install
# requires Node 20+ git clone github.com/alcabal-1/zentient-regen npm install && node ./cli.js evaluate "…"
Read the source
Open repo on GitHub

Full system prompt, 10-action sample dataset, worked example, OpenClaw SOUL.md, architecture diagram.

github.com/alcabal-1/zentient-regen →

Empty lot → community food forest

A first-cell candidate. Lot-scale (~0.5 acre), 500-household catchment. Pre-feasibility — site, tenure, stakeholder alignment, and funding all unconfirmed. What the bot returns:

INPUT

"We want to turn an empty lot into a community food forest with solar, compost, and youth programming."

Impact Map (excerpt)

  • Solar generation (kWh/year) provable Direct measurement once installed and metered.
  • Compost throughput (tons/year diverted from landfill) provable Intake + output logs; municipal diversion verification optional.
  • Local food production (~X lbs/year) plausible Food-forest yields highly variable; depend on plant selection + 3–5 year establishment + maintenance.
  • Stormwater management / impervious-surface reduction plausible Standard green-infrastructure benefit; lot conversion measurably increases permeability.
  • Biodiversity uplift plausible Native-plant + pollinator habitat — well-established benefit (Drawdown § Land Sinks).
  • Carbon sequestration / ReFi credit potential speculative Soil-carbon protocols exist (Verra, Puro.earth) but lot-scale projects rarely meet additionality + permanence + verification cost thresholds.
  • Youth empowerment / educational outcomes not yet verifiable "Empowerment" lacks operational definition. Reformulate as participation hours, demonstrated skills, longitudinal engagement.
  • Community cohesion / mutual aid not yet verifiable Real benefit, hard to measure. Sense of Community Index (Chavis & Wandersman); pre/post surveys.

7-Day Action Plan (excerpt)

DayActionOwnerDeliverable
1 Title search + tax-default status check on lot Operator + zentient-permit 1-page status memo
1 Zoning lookup + community-garden ordinance scan zentient-permit Memo + ordinance excerpt
2 Cooperative Extension contact — soil-health + Phase I ESA referral Operator Email sent + meeting requested
3–4 Door-to-door survey deployment — aim ≥ 30% household reach Community organizer + 2 volunteers Response log
4 Fiscal sponsor outreach (Earth Island / Tides / local options) Operator 3 outreach emails sent
6 Solar feasibility pre-screen — PVWatts + utility consult zentient-PTO Pre-screen memo
7 Week-1 retrospective + Week-2 plan Operator + zentient-regen Updated proof plan

Route-to-Swarm

  • invoke_skill("zentient-permit", "research_zoning", { lot_address, jurisdiction })
  • invoke_skill("zentient-google", "create_calendar_event", { stakeholder_meeting_topic, attendees })
  • invoke_skill("zentient-field", "schedule_site_visit", { lot_address, baseline_observations })
  • invoke_skill("zentient-PTO", "solar_pre_screen", { lot_address, lot_size_sqft })
  • ask_superorganism("similar food-forest projects in [region] last 3 years")

Full output: classification, partner-fit grant programs, top risks, claim labels glossary, GLM regional precedent scan — examples/proof-plan-01.md →

Three failure modes

Public-goods coordination has well-studied failure modes. Zentient Regen is infrastructure to address them.

1
Unverified impact claims destroy credibility
"Regenerative" and "carbon-positive" claims collapse under scrutiny when not backed by evidence. The ElizaOS lesson: speculation without verified work erodes trust across the whole sector.
Default to plausible/speculative until evidence ships.
2
Coordination overhead crushes volunteer-led projects
Community-led regenerative projects fail to organizer burnout more than 60% of the time. The bottleneck isn't ideas — it's the routing, the partner introductions, the proof plans that nobody has time to write.
Pretty 2003; Putnam 2000.
3
No reusable proof-plan infrastructure exists
Every project re-discovers verification, funding fit, and partner routing from scratch. Zentient Regen is open-source so any DAO, community group, or land trust can deploy it and contribute back.
Apache 2.0 · alcabal-1/zentient-regen

Reasoning, knowledge, routing

The bot drafts and routes. Humans decide. The agent does not autonomously commit funding, sign contracts, or execute on real-world projects. Output is structured for human review, refinement, and execution.

Input
User opportunity in plain language
Reasoning
Claude Sonnet 4.6 orchestrator
System prompt encodes proof discipline + claim labeling
Knowledge
Action knowledge base
10 sample actions (Apache 2.0) · Drawdown / SDG / Gallup citations · full 50+ proprietary substrate under commercial license
Optional
Z.AI GLM-5.1 regional scan
Peripheral — gracefully stubs offline
Output
6-output proof plan
Classification · impact map · proof plan · routing · 7-day plan · claim discipline
MODE 01
Standalone CLI
This repository — today's hackathon ship. Local execution, markdown output.
MODE 02
Superorganism node
WebSocket + MCP into the Zentient Superorganism — peer agents discover and invoke skills.
MODE 03
OpenClaw runtime
Plugin manifest. Same agent, OpenClaw-managed lifecycle.

What's open. What's commercial.

The repository is a real, runnable, useful demo engine — sufficient to evaluate the proof-plan method and to deploy basic regenerative-opportunity workflows. The full commercial product carries the proprietary IP that distinguishes operator deployments.

Open Source · Apache 2.0

Engine + sample dataset

  • Engine source code
  • 10-action sample dataset (Drawdown / SDG / Gallup)
  • Markdown output format
  • OpenClaw alignment + SOUL.md
  • Worked food-forest example
  • Standalone CLI deployment mode
View repo →
Commercial license

Full Ecoterium / Zentient substrate

  • Full regenerative-action library (50+ actions, Dodecahedron of the Infinity Economy)
  • Sacred-dynamics coordination framework
  • Maslow-Myer cone full matching engine
  • Sentient Habitat orchestration (neighborhood-scale)
  • Hosted deployment + proof-certificate audit trails
  • Partner / supplier / funder evaluation rubrics
  • Triple-helix doctrine + autoresearcher concept
Commercial inquiry →

2026 BETA Hackathon

Berkeley Emerging Technology Association · Frontier Tower · April 26 2026

Built at the 2026 BETA Hackathon. Submitted to BETA Open Innovation, GCC OpenClaw Grants — Impact Evaluation Track, and the Z.AI sponsor track (uses GLM-5.1).

Zentient and Zentient Regen are reserved marks of Alvaro Cabal. Open-source code is licensed under Apache 2.0. Proprietary substrate — sacred-dynamics, Maslow-Myer matching, Sentient Habitat orchestration — is the intellectual property of Alvaro Cabal / Zentient and available under commercial license.