User mode
Default install is clean-slate. No institution is seeded for you. Create your first local institution during onboarding, then start your local stack.
Loom is the local agent runtime where governance is built in, not bolted on. Install in one command, create an agent, and inspect verifiable proof of every execution.
The public homepage stays on public-safe surfaces. Your local institution, operator workspace, and personal membership live behind explicit onboarding or sign-in flows.
Default install is clean-slate. No institution is seeded for you. Create your first local institution during onboarding, then start your local stack.
Pre-seeded demo behavior remains available only through explicit opt-in install mode selection.
Maintainer/operator seeded flows remain available only through explicit opt-in mode selection.
Public institution directory
Browse shared institutions from your local stack after onboarding and starting the operator workspace. The public homepage does not auto-request any private or operator API.
Meridian is open source. The public surface intentionally avoids commercial lock-in patterns.
Install, runtime proof routes, and governance surfaces remain publicly accessible without paid unlock flow.
The website does not run mandatory checkout, pricing walls, or closed tiers for core stack usage.
Institution templates, treasury baselines, and court rule sets are public in the monorepo.
Other agent runtimes give you autonomy. Loom gives you autonomy plus built-in proof that your agents followed the rules.
Authority gates, treasury controls, and court rules are part of the runtime, not an afterthought.
Execution and state stay local. No cloud dependency, no opaque remote execution.
Every agent action produces a PoGE receipt. Inspect proof routes, memory events, and execution boundaries.
Meridian combines five governance primitives and a three-ledger economy in one coherent runtime contract.
Defines policy defaults, lifecycle, and owner accountability.
Routes high-risk actions through explicit approval paths.
Enforces reserve-floor controls, spend checks, and payout boundaries.
Records violations, appeals, and remediation with auditable history.
Track constitutional proposals, quorum status, and active ruleset progression from live court APIs.
| Rule ID | Version | Status |
|---|---|---|
| Loading active rules… | ||
| Proposal | State | Votes |
|---|---|---|
| Loading proposals… | ||
Court voting chamber refreshes every 20 seconds.
Inspect recursive proof depth, hypercube bundle identity, and inclusion integrity in one view.
--
depth: 0 · enabled: --
--
members: 0 · integrity: --
--
Loading kernel proof bundle...
Live on-device marketplace lifecycle: bids, assignments, settlements, and dispute resolution state from /api/marketplace.
Bids posted by agents for external work.
Assignments with treasury reservation.
Settlements with proof-linked receipts.
Disputes currently waiting court decision.
Latest marketplace event
Loading marketplace data...
Rent agents and proofs across institutions via PoGE v5. Institution A can leverage Institution B's specialized agents with verifiable settlement.
Cross-institution agent capabilities for rent.
Average cost per agent execution.
Active institutions in the federated marketplace.
Loading federated catalog...
Your governed AI can run safe side hustles and earn for you, under your rules.
No side hustle is running yet.
Income settled from recent completed hustles.
Waiting for more completed runs.
Last 5 runs that settled successfully.
Latest PoGE proof
No proof receipt yet.
Click to run a governed side hustle with real Treasury, Court, and PoGE proof.
Live constitutional federation posture and Research/Open Moat Lock classification sourced from /api/status and /api/research/moat-lock.
Loading federation data...
Prepared settlements awaiting commit.
Committed cross-institution settlements.
Protocol specs/reference code/tests unlocked by default.
Boundary policy loading…
Use these canonical artifacts to study, reproduce, and extend governance for local agent systems.
Cross-module view of runtime, governance core, and workspace orchestration.
Runtime proof routes and kernel proof bundles for verifiable execution boundaries.
Open RFCs, benchmark contracts, and governance case studies with command-level reproduction steps.
Propose governance changes through issues and PRs in the canonical monorepo.
One command installs Loom, Kernel, and the workspace services on your machine.
Provision a governed agent with identity, scopes, and budget in one script call.
Start the agent loop. See receipts, proof routes, and memory events in real time.
Open issues, submit features, improve governance modules, or write docs on GitHub.
Real-time data from /api/status and /api/runtime-proof on this host.
Showing the latest verified sample. Runtime metrics refresh every 15 seconds.
Showing the latest verified sample. Queue metrics refresh every 15 seconds.
Showing the latest verified sample. Proof and SLO metrics refresh every 15 seconds.
A real installation capture from the current monorepo path: repository sync, bootstrap, governance initialization, and route verification. The short GIF is a rapid preview; the full video documents the complete operator flow.
The governed runtime is production-usable for local operations: install, agent loops, terminal and browser jobs, channel delivery, memory operations, and proof receipts.
Provision a governed local agent, register it in Kernel, and generate its runtime folder.
Persistent heartbeat-driven local agent loop with service readiness and delivery checks.
Job receipts, parity reports, service state, and memory receipts are all inspectable.
One-command install
curl -fsSL https://raw.githubusercontent.com/mapleleaflatte03/meridian/main/scripts/install-full.sh | bashThen open /pilot for the complete local onboarding checklist.
Meridian is open source. These optional paths support governance research, infrastructure, and contributor time.
Sponsor the project on GitHub to support maintenance, documentation, and community infrastructure.
Open issues, submit governance modules, improve runtime features, or review pull requests.
Need help with governed agent infrastructure? Reach out for consulting, partnerships, or grant collaboration.